WO2022062459A1 - Automatic measurement method and device for fetal structural characteristic - Google Patents

Automatic measurement method and device for fetal structural characteristic Download PDF

Info

Publication number
WO2022062459A1
WO2022062459A1 PCT/CN2021/096822 CN2021096822W WO2022062459A1 WO 2022062459 A1 WO2022062459 A1 WO 2022062459A1 CN 2021096822 W CN2021096822 W CN 2021096822W WO 2022062459 A1 WO2022062459 A1 WO 2022062459A1
Authority
WO
WIPO (PCT)
Prior art keywords
structural
feature
structural features
contour
structural feature
Prior art date
Application number
PCT/CN2021/096822
Other languages
French (fr)
Chinese (zh)
Inventor
谢红宁
汪南
冼建波
梁喆
吴海涛
Original Assignee
广州爱孕记信息科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州爱孕记信息科技有限公司 filed Critical 广州爱孕记信息科技有限公司
Publication of WO2022062459A1 publication Critical patent/WO2022062459A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
    • A61B5/1072—Measuring physical dimensions, e.g. size of the entire body or parts thereof measuring distances on the body, e.g. measuring length, height or thickness
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/107—Measuring physical dimensions, e.g. size of the entire body or parts thereof
    • A61B5/1079—Measuring physical dimensions, e.g. size of the entire body or parts thereof using optical or photographic means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30044—Fetus; Embryo

Definitions

  • the present invention relates to the field of image technology, and in particular, to an automatic measurement method and device for fetal structural features.
  • the acquisition method of fetal growth and development is as follows: the examiner (for example, a medical staff) manually determines the long axis and short axis of the fetal head circumference or abdominal circumference by moving the trackball, and obtains the oval shape of the fetal head circumference or abdominal circumference. , and then calculate the fetal head circumference or abdominal circumference through the ellipse circumference formula, so as to obtain the measurement result of the fetal head circumference or abdominal circumference, and determine the growth and development of the fetus according to the measurement result.
  • the examiner for example, a medical staff
  • calculate the fetal head circumference or abdominal circumference through the ellipse circumference formula so as to obtain the measurement result of the fetal head circumference or abdominal circumference, and determine the growth and development of the fetus according to the measurement result.
  • the technical problem to be solved by the present invention is to provide an automatic measurement method and device for fetal structural characteristics, which can obtain high-precision fetal measurement results, so as to accurately determine the growth and development of the fetus.
  • the first aspect of the present invention discloses an automatic measurement method for fetal structural characteristics, the method comprising:
  • the acquired fetal ultrasound image is input into the determined feature detection model for analysis, and the analysis result output by the feature detection model is obtained as the feature information of the fetal ultrasound image, and the feature information of the fetal ultrasound image includes the a category of at least one structural feature of the fetal ultrasound image and a contour of at least one of the structural features;
  • a measurement method that matches the contour of the structural feature is determined, and a measurement operation is performed on the contour of the structural feature according to the measurement method of each of the structural features to obtain the geometric parameters of the structural feature.
  • the method further includes:
  • a measurement operation is performed on the contour of the feature to obtain the geometric parameters of the structural feature.
  • the contour of each structural feature is the contour of the structural feature after preprocessing, and the preprocessing operation includes a smoothing operation.
  • the method further includes:
  • the preprocessing operation is an operation of obtaining the contours of the preprocessed structural features, where all the structural features include all the target structural features.
  • the outline of each of the structural features corresponds to a plurality of nodes
  • all the structural features include at least one of the structural features of the cranial brain, the structural features of the abdominal circumference, and other structural features, and the other structural features include the structure of the cerebellum At least one of features, structural features of the third ventricle, structural features of the lateral ventricle, structural features of the septum pellucidum, structural features of the humerus, and structural features of the femur.
  • the measurement method that matches the outline of each structural feature is determined according to the category of each structural feature, and The measurement method performs a measurement operation on the outline of the structural feature to obtain the geometric parameters of the structural feature, including:
  • the geometric parameters corresponding to the cranial structural feature are determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature.
  • the contour of the brain structure feature includes the outer contour and the inner contour;
  • the structural feature is the abdominal circumference structural feature
  • the structural feature is the other structural feature
  • fit the contour of each other structural feature based on the determined graph fitting method obtain the target node of the other structural feature, and based on each of the other structural features
  • the target node of the feature determines the geometric parameters of the other structural feature
  • the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
  • the contour of each of the other structural features is fitted based on the determined graph fitting method, to obtain the target node of the other structural feature, and based on For each target node of the other structural feature, determine the geometric parameters of the other structural feature, including:
  • the graph fitting mode is the minimum circumscribed rectangle fitting mode
  • the contours of the other structural features are fitted based on the determined circumscribed rectangle.
  • the contour of the structural feature has a target intersection
  • stop fitting the contour of the other structural feature obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the sweep of the other structural feature according to the category of the other structural feature. check the direction;
  • the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to indicate that the scanning direction is parallel to the short side of the minimum circumscribed rectangle.
  • the contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle.
  • the contour of each of the other structural features is fitted based on the determined graph fitting method, to obtain the target node of the other structural feature, and based on For each target node of the other structural feature, determine the geometric parameters of the other structural feature, including:
  • the graph fitting method is the ellipse fitting method
  • the contours of the other structural features are fitted based on the determined ellipse fitting algorithm, and the fitted contours of the other structural features are obtained.
  • the shape of the outline of the other structural features is an ellipse
  • the target scanning direction includes one of a first target scanning direction or a second target scanning direction, wherein the first target scanning direction is used to represent a direction from parallel to the fitted
  • the short axis of the contour of the other structural feature scans the contour of the other structural feature
  • the second target scanning direction is used to indicate that the long axis of the contour of the other structural feature is parallel to the fitting. Scan the contours of other structural features described above.
  • the geometry corresponding to the cranial structural feature is determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature parameters, including:
  • the biparietal diameter geometric parameter corresponding to the cranial brain structural feature is determined.
  • a second aspect of the present invention discloses an automatic measuring device for fetal structural features, the device comprising:
  • An analysis module which inputs the acquired fetal ultrasound image into the determined feature detection model for analysis
  • the acquisition module is configured to acquire the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the category of at least one structural feature of the fetal ultrasound image and at least one of the features of the fetal ultrasound image. outline of structural features;
  • the measurement module is configured to determine a measurement mode that matches the contour of the structural feature according to the category of each of the structural features, and perform a measurement operation on the contour of the structural feature according to the measurement mode of each of the structural features to obtain the Geometric parameters of structural features.
  • the device further includes:
  • the preprocessing module is configured to, after the acquisition module acquires the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, executes on the contour of each of the structural features based on the determined preprocessing method Preprocessing operation, obtaining the contour of the preprocessed structural feature, and triggering the measurement module to perform the measurement operation on the contour of the structural feature according to the measurement method of each of the structural features, to obtain the geometry of the structural feature
  • the contour of each structural feature is the contour of the preprocessed structural feature
  • the preprocessing operation includes a preprocessing operation.
  • the device further includes:
  • the judgment module is used for, after the acquisition module obtains the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, judges the contour of each of the structural features according to the feature information of the fetal ultrasound image Whether the determined preprocessing condition is met, when the judgment result is no, trigger the measurement module to perform the measurement operation for the contour of the structural feature according to the measurement method of each of the structural features, and obtain the profile of the structural feature. manipulation of geometric parameters;
  • the screening module is used for screening all the target structural features of the preprocessing condition from all the structural features when the judgment result of the judgment module is yes, and triggering the preprocessing module to perform the
  • the preprocessing method is an operation of performing a preprocessing operation on the contour of each of the structural features to obtain the contours of the preprocessed structural features, and all the structural features include all the target structural features.
  • the outline of each of the structural features corresponds to a plurality of nodes
  • the preprocessing module performs a preprocessing operation on the contour of each structural feature based on the determined preprocessing method, and the method for obtaining the contour of the preprocessed structural feature is specifically:
  • all the structural features include at least one of the structural features of the cranial brain, the structural features of the abdominal circumference, and other structural features, and the other structural features include the structure of the cerebellum At least one of features, structural features of the third ventricle, structural features of the lateral ventricle, structural features of the septum pellucidum, structural features of the humerus, and structural features of the femur.
  • the measurement module determines, according to the category of each of the structural features, a measurement method that matches the outline of the structural feature, and according to each of the structural features
  • the measurement method of the feature performs a measurement operation on the outline of the structural feature, and the method for obtaining the geometric parameters of the structural feature is as follows:
  • the geometric parameters corresponding to the cranial structural feature are determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature.
  • the contour of the brain structure feature includes the outer contour and the inner contour;
  • the structural feature is the abdominal circumference structural feature
  • the structural feature is the other structural feature
  • fit the contour of each other structural feature based on the determined graph fitting method obtain the target node of the other structural feature, and based on each of the other structural features
  • the target node of the feature determines the geometric parameters of the other structural feature
  • the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
  • the measurement module fits the contour of each of the other structural features based on the determined graphic fitting method, and obtains the target node of the other structural feature, And based on the target node of each of the other structural features, the specific way of determining the geometric parameters of the other structural features is:
  • the graph fitting mode is the minimum circumscribed rectangle fitting mode
  • the contours of the other structural features are fitted based on the determined circumscribed rectangle.
  • the contour of the structural feature has a target intersection
  • stop fitting the contour of the other structural feature obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the sweep of the other structural feature according to the category of the other structural feature. check the direction;
  • the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to indicate that the scanning direction is parallel to the short side of the minimum circumscribed rectangle.
  • the contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle.
  • the measurement module fits the contour of each of the other structural features based on the determined graphic fitting method, and obtains the target node of the other structural feature, And based on the target node of each of the other structural features, the specific way of determining the geometric parameters of the other structural features is:
  • the graph fitting method is the ellipse fitting method
  • the contours of the other structural features are fitted based on the determined ellipse fitting algorithm, and the fitted contours of the other structural features are obtained.
  • the shape of the outline of the other structural features is an ellipse
  • the target scanning direction includes one of a first target scanning direction or a second target scanning direction, wherein the first target scanning direction is used to represent a direction from parallel to the fitted
  • the short axis of the contour of the other structural feature scans the contour of the other structural feature
  • the second target scanning direction is used to indicate that the long axis of the contour of the other structural feature is parallel to the fitting. Scan the contours of other structural features described above.
  • the measurement module determines the corresponding cranial structure feature based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature
  • the way of the geometric parameters is as follows:
  • the biparietal diameter geometric parameter corresponding to the cranial brain structural feature is determined.
  • a third aspect of the present invention discloses another automatic measuring device for fetal structural features, the device comprising:
  • a processor coupled to the memory
  • the processor invokes the executable program code stored in the memory to execute the automatic measurement method for fetal structural features disclosed in the first aspect of the present invention.
  • a fourth aspect of the present invention discloses a computer storage medium, the computer storage medium stores computer instructions, and when the computer instructions are invoked, is used to execute the automatic measurement method for fetal structural features disclosed in the first aspect of the present invention.
  • the acquired fetal ultrasound image is input into the determined feature detection model for analysis, and the analysis result output by the feature detection model is obtained as the feature information of the fetal ultrasound image, the feature information of the fetal ultrasound image Including the category of at least one structural feature of the fetal ultrasound image and the contour of at least one structural feature; determining a measurement mode matching the contour of the structural feature according to the category of each structural feature, and according to the measurement mode of each structural feature A measurement operation is performed on the contour of the structural feature to obtain the geometrical parameters of the structural feature.
  • the implementation of the present invention can quickly obtain high-precision fetal structural features by automatically acquiring the categories and contours of the structural features of the fetus, and measuring the contours of the structural features according to the corresponding measurement methods, without manually obtaining the geometric parameters of the fetal structural features.
  • the measurement results of the fetus can be accurately determined, so as to accurately determine the growth and development of the fetus; and by inputting the fetal ultrasound image into the feature detection model, the acquisition efficiency of the measurement results of the fetal structural features can be improved.
  • FIG. 1 is a schematic flowchart of an automatic measurement method for fetal structural features disclosed in an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of another automatic measurement method for fetal structural features disclosed in an embodiment of the present invention
  • FIG. 3 is a schematic structural diagram of an automatic measuring device for fetal structural features disclosed in an embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention.
  • the invention discloses an automatic measuring method and device for fetal structural features, which can automatically obtain the category and contour of fetal structural features, and measure the contours of structural features according to corresponding measurement methods, without manually obtaining the geometrical features of fetal structural features. parameters, can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, the acquisition efficiency of fetal structural features can be improved. Each of them will be described in detail below.
  • FIG. 1 is a schematic flowchart of an automatic measurement method for fetal structural features disclosed in an embodiment of the present invention. 1 can be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention.
  • the automatic measurement method of fetal structural features may include the following operations:
  • the fetal ultrasound image may be a single frame picture or a dynamic image.
  • the fetal ultrasound image when the fetal ultrasound image is a single-frame picture, the fetal ultrasound image can be continuously input into the feature detection model for analysis according to the predetermined frame rate, so that by inputting the continuous multiple frames of fetal ultrasound images into the feature detection model for analysis, It is beneficial to carry out multiple analysis of the same structural feature, thereby facilitating multiple measurement of the contour of the same structural feature, thereby improving the measurement accuracy of the structural feature of the fetal ultrasound image, thereby improving the accuracy of determining the growth and development of the fetus;
  • the measurement operation can also be performed on the contours of more structural features of the fetal ultrasound image, thereby facilitating determining the growth and development of the fetus corresponding to the fetal ultrasound image according to geometric parameters of different structural features.
  • the predetermined frame rate is related to the structural characteristics of the fetal ultrasound image to be measured, that is, the frame rate is selected according to the structural characteristics of the fetal ultrasound image to be measured, for example, if the abdominal circumference structure needs to be measured. feature, the frame rate can be 30 frames/second; if the four-chamber structural feature needs to be measured, the frame rate can be 60 frames/second. In this way, the corresponding frame rate is selected according to the structural features of the fetal ultrasound image to be measured, which is beneficial to improve the measurement efficiency and accuracy of the geometric parameters of the structural features of the fetal ultrasound image.
  • the feature detection model may first divide the fetal ultrasound image into multiple frames of fetal ultrasound pictures before analyzing the feature information of the fetal ultrasound image. In this way, by performing the frame segmentation operation on the dynamic fetal ultrasound image, the structural features of the fetal ultrasound image can be subdivided, thereby helping to improve the measurement accuracy of the geometric parameters of the structural features of the fetal ultrasound image.
  • the feature information of the fetal ultrasound image may further include at least one of graphic coordinates, size, and position probability of the structural feature of the fetal ultrasound image, which is not limited in the embodiment of the present invention.
  • the graphic coordinates may include polygon coordinates or elliptical coordinates, wherein the polygon coordinates may include odd polygon coordinates or even polygon coordinates, such as: pentagonal coordinates, rectangular coordinates, and the selection of polygon coordinates depends on the shape of the structural feature, which can improve the structure The coordinate acquisition accuracy of the feature.
  • the feature information of the fetal ultrasound image also includes polygonal contour information of the structural features of the fetal ultrasound image, such as polygonal contour coordinates, so that the more content the feature information of the fetal ultrasound image includes, the more conducive to improving the growth and development of the fetus. The accuracy of the determination of the situation.
  • each frame of fetal ultrasound image has a unique corresponding frame sequence number, so that by setting a unique frame sequence number for each frame of fetal ultrasound image, it is possible to measure the structural features of the fetal ultrasound image during the measurement process of the fetal ultrasound image. , clearly distinguish each frame of fetal ultrasound images and facilitate the management of fetal ultrasound images and related information (eg, geometric parameters) of their structural features.
  • all structural features may include at least one of cranial brain structural features, abdominal circumference structural features, and other structural features, and other structural features include cerebellar structural features, third ventricle structural features, lateral ventricle structural features, transparency At least one of the structural features of the septum, the structural features of the humerus, and the structural features of the femur is not limited in the embodiment of the present invention.
  • the contour of each structural feature has a corresponding measurement method. It should be noted that, whether the acquired structural features are normal structural features or abnormal structural features, the corresponding measurement methods are the same.
  • the automatic measurement method of fetal structural features described in FIG. 1 can automatically obtain the category and contour of fetal structural features, and measure the contour of structural features according to the corresponding measurement method, without manually obtaining the geometric parameters of fetal structural features. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results.
  • a measurement method that matches the contour of each structural feature is determined according to the category of each structural feature, and a measurement operation is performed on the contour of the structural feature according to the measurement method of each structural feature to obtain the Geometric parameters of structural features, which can include:
  • the geometric parameters corresponding to the cranial structure feature are determined based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature, and the outline of the cranial structure feature includes the cranial structure feature The outer contour of the brain and the inner contour of the structural features of the brain;
  • the perimeter corresponding to the contour of the abdominal circumference structural feature is obtained as the geometric parameter of the abdominal circumference structural feature
  • the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
  • the contour of each other structural feature is fitted based on the determined graph fitting method, and the target node of the other structural feature is obtained, And based on the target node of each other structural feature, determine the geometric parameters of the other structural feature, including: fitting the contour of the other structural feature based on the determined circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle is consistent with other structural features.
  • the contour of the feature has a target intersection
  • stop fitting the contour of other structural features obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the scanning direction of the other structural feature according to the category of the other structural feature, and based on the scanning
  • the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to represent a pair of short sides parallel to the smallest circumscribed rectangle.
  • the contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle, for example: scanning of the structural feature of lateral ventricle enlargement
  • the direction is parallel to the short side of the smallest circumscribed rectangle, and the scanning direction of the structural features of the left atrium is parallel to the long side of the smallest circumscribed rectangle.
  • the target intersection point is in the process of wirelessly approximating the contours of other structural features by the circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, stop fitting other structural features.
  • the outline of the structural feature can improve the accuracy of determining the scanning direction of the structural feature, thereby further improving the accuracy, reliability and efficiency of obtaining geometric parameters of the structural feature.
  • this optional embodiment can obtain the geometric parameters of the structural features by fitting the structural features through the minimum circumscribed rectangle, and select different scanning directions for different types of structural features, which can improve the accuracy of the geometric parameters of the structural features. performance, reliability and efficiency.
  • the geometric parameters of the other structural features may also be obtained by: The vertical line is obtained, and the two intersection points of the vertical line and the outline of other structural features, as well as the line segment formed by connecting the two intersection points, are obtained as the geometric parameters of other structural features.
  • the acquisition methods of the geometric parameters of other structural features include the above two acquisition methods, the mean value of the geometric parameters of the other structural features obtained by the two acquisition methods is acquired as the final transverse diameter geometric parameters of the other structural features.
  • the acquisition methods of the cerebellar transverse diameter include the above two acquisition methods
  • the average value of the cerebellar transverse diameters obtained by the two acquisition methods is obtained as the final transverse diameter of the cerebellum structural features, thereby improving the geometric parameters of other structural features. accuracy and reliability of acquisition.
  • the contour of each other structural feature is fitted based on the determined graph fitting method to obtain the target node of the other structural feature, and based on the target node of each other structural feature,
  • the specific way of determining the geometric parameters of the other structural features is as follows:
  • the graph fitting method is the ellipse fitting method
  • the contours of other structural features are fitted based on the determined ellipse fitting algorithm to obtain the contours of the fitted other structural features and the shape of the fitted contours of the other structural features. is oval;
  • the target scanning direction includes one of the first target scanning direction or the second target scanning direction, wherein the first target scanning direction is used to represent the direction from the contour parallel to the other structural features after fitting.
  • the short axis scans the contours of other structural features
  • the second target scanning direction is used to indicate that the contours of other structural features are scanned from the long axis parallel to the contours of the other structural features after fitting.
  • the scanning direction of the cranial structural feature is parallel to the short axis of the contour of the cranial structural feature after ellipse fitting, and the scanning line segment with the largest distance value is taken as the transverse diameter of the cranial structural feature.
  • the scanning direction of other structural features can also be determined based on the minimum circumscribed rectangle fitting method and the ellipse fitting method, for example, the scanning direction of the cerebellar transverse diameter.
  • the direction of the search can be determined by the minimum circumscribed rectangle and the elliptical halo structure of the skull.
  • the scanning direction of the structural feature is determined jointly by the ellipse fitting method and the minimum circumscribed rectangle fitting method, which can improve the accuracy and reliability of the scanning direction of the structural feature, thereby improving the accuracy of determining the geometric parameters of the structural feature. and reliability.
  • this optional embodiment measures the geometric parameters of the corresponding structural features by adopting different geometric parameter measurement methods according to different structural features, which can not only achieve the acquisition of the geometric parameters of the structural features of the fetal ultrasound image, but also can improve the fetal ultrasound image.
  • determining the geometric parameters corresponding to the cranial structure feature may include:
  • the second intersection point is to determine the geometric parameters of the biparietal diameter corresponding to the structural features of the skull.
  • the calculation method of the head circumference geometric parameter corresponding to the cranial structure feature is as follows:
  • C is the head circumference geometric parameter of the cranial structure feature, namely the head circumference;
  • C 1 is the second circumference of the outer contour of the cranial structure feature;
  • C 2 is the first circumference of the inner contour of the cranial structure feature long.
  • the first intersection includes a first sub-intersection and a second sub-intersection
  • the second intersection includes a third sub-intersection and a fourth sub-intersection.
  • the distance between the first sub-intersection and the third sub-intersection is smaller than the distance from the fourth sub-intersection.
  • the biparietal diameter geometric parameter corresponding to the cranial structure feature specifically: connect the first line segment formed by the first sub-intersection and the fourth sub-intersection as the cranial structure feature
  • the corresponding biparietal diameter geometric parameter that is, the biparietal diameter length; or, the second line segment formed by connecting the second sub-intersection and the third sub-intersection, as the biparietal diameter geometric parameter corresponding to the structural features of the brain; or, obtain the first
  • the mean value of the first line segment and the second line segment is used as the geometric parameter of the biparietal diameter corresponding to the structural features of the brain.
  • the perimeter of the inner and outer contours of the structural features of the skull, and the intersection of the midline corresponding to the midline of the brain and the inner and outer contours of the structural features of the skull can realize the structural features of the skull.
  • the perimeter of the head circumference and the length of the double parietal diameter were obtained.
  • the method may further include the following steps:
  • the gestational age of the fetus corresponding to the fetal ultrasound image is determined according to the geometric parameter of at least one structural feature of the fetal ultrasound image.
  • the gestational age of the fetus corresponding to the fetal ultrasound image may be determined independently according to the geometric parameters of each structural feature of the fetal ultrasound image, or the fetal ultrasound image may be jointly determined according to the geometric parameters of multiple structural features The corresponding gestational age of the fetus.
  • the gestational age of the fetus corresponding to the fetal ultrasound image can be determined individually or jointly according to the geometric parameters of the structural features of the fetal ultrasound image, thereby improving The accuracy of determination of fetal gestational age.
  • the method may further include the following steps:
  • all the structural features of the samples include the structural features of the skull of the sample, the structural features of the abdominal circumference of the sample, the structural features of the cerebellum of the sample, the structural features of the third ventricle of the sample, the structural features of the lateral ventricle of the sample, the structural At least one of the structural features of the humerus and the structural features of the sample femur is not limited to this optional embodiment.
  • each sample structure feature includes a positive sample structure feature and a negative sample structure feature, wherein the positive sample structure feature is a normal sample structure feature and the negative sample structure feature is an abnormal sample structure feature.
  • the initial feature detection model may include a target detection model, an instance segmentation model, a semantic segmentation model, and a combination of one or more models, which are not limited in this optional embodiment.
  • the trained feature detection model corresponds to the initial feature detection model, that is, the trained feature detection model may include the trained target detection model, the trained instance segmentation model, and the trained semantic segmentation model, etc.
  • the optional embodiment is not limited to a combination of one or more models in the feature information of the fetal ultrasound image.
  • different sample structural features correspond to different measurement elements, specifically: the measurement elements corresponding to the sample cerebellar structural features are the lines connecting the widest positions of the sample cerebellum structural features (cerebellar hemispheres), and the connecting line Perpendicular to the midline of the brain; the measurement element corresponding to the structural features of the third ventricle of the sample is the line segment of the widest region of the third ventricle cavity, and the line segment is perpendicular to the midline of the brain; the measurement element corresponding to the structural features of the lateral ventricle of the sample is the lateral ventricle cavity The line segment of the widest region, and the line segment is perpendicular to the longest side of the circumscribed rectangle; the measurement element corresponding to the transparent septum structure feature of the sample is the line segment of the widest region of the transparent septum cavity, and the line segment is perpendicular to the longest side of the circumscribed rectangle; the sample humerus The measurement element corresponding to the structural feature is the line segment of the
  • this optional embodiment can facilitate the subsequent use of the feature detection model by pre-training a feature detection model that conforms to this scheme, thereby improving the information of the structural features of the fetal ultrasound image, and then improving the geometrical geometry of the structural features of the fetal ultrasound image.
  • the efficiency and accuracy of parameter acquisition; and in the process of training the feature detection model, the measurement elements corresponding to structural features are trained together, which can improve the training accuracy and reliability of the feature detection model, which is conducive to further improving fetal ultrasound images.
  • FIG. 2 is a schematic flowchart of another method for automatically measuring fetal structural features disclosed in an embodiment of the present invention.
  • the automatic measurement method for fetal structural features described in FIG. 2 may be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention.
  • the automatic measurement method of fetal structural features may include the following operations:
  • the preprocessing operation includes a smoothing operation. Further, the preprocessing operation may further include at least one of image enhancement, image filtering, and image segmentation operations, which are not limited in this optional embodiment.
  • step 202 may occur after step 203, or before step 203, or may occur at the same time as step 203, which is not implemented in this embodiment of the present invention. limited.
  • the embodiment of the present invention further performs a preprocessing operation on the contour of the structural feature of the fetal ultrasound image, such as a smoothing operation, which can filter out the noise in the contour of the structural feature, A clear outline of the structural feature is obtained, thereby helping to improve the efficiency and accuracy of obtaining the geometrical parameters of the structural feature.
  • step 204 Perform a measurement operation on the contour of the structural feature according to the measurement method of each structural feature to obtain the geometric parameters of the structural feature.
  • the contour of each structural feature in step 204 is the preprocessed structural feature. contour.
  • step 201 for other descriptions of step 201, step 203, and step 204, please refer to the detailed description in Embodiment 1, and details are not repeated in this embodiment of the present invention.
  • the method may further include the following steps:
  • step 202 When it is determined that there are contours that meet the preprocessing conditions in the contours of all the structural features, all target structural features that meet the preprocessing conditions are screened from all the structural features, and the execution of step 202 is triggered. At this time, all the structures in step 202 Features include all target structural features. It should be noted that when there is a target structural feature, that is, there is a structural feature that needs to be preprocessed, the operation of determining the matching measurement method can be directly performed on the contours of other structural features that do not need to be preprocessed. After the preprocessing operation is performed on the target structure feature, the operation of determining a matching measurement mode may be performed together with the preprocessed target structure feature, which is not limited in this optional embodiment.
  • the preprocessing condition if the preprocessing condition is not met, the measurement operation of the contour of the structural feature is directly performed. , which can improve the measurement efficiency of the geometric parameters of structural features, thereby improving the efficiency of determining the growth and development of the fetus; if the preprocessing conditions are met, the preprocessing operation is performed on the contours of the structural features that meet the preprocessing conditions, and accurate and reliable information can be obtained.
  • the contour of the structural feature improves the measurement accuracy and reliability of the geometric parameters of the structural feature, thereby further improving the accuracy and reliability of the determination of the growth and development of the fetus.
  • judging whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image may include:
  • each structural feature it is determined whether the structural feature is the determined category of the structural feature that needs to perform a preprocessing operation, and when the judgment result is yes, it is determined that the structural feature satisfies the determined preprocessing condition; or,
  • each structural feature it is determined whether the contour of the structural feature satisfies the determined parameter measurement condition, and when the judgment result is no, it is determined that the structural feature satisfies the determined preprocessing condition.
  • the category of each structural feature is the category of the structural feature that needs to perform a preprocessing operation and it is determined that the contour of the structural feature does not meet the determined parameter measurement conditions. It is determined that the structural feature satisfies the determined preprocessing conditions. This can not only further improve the accuracy of determining that the contour of the structural feature of the fetal ultrasound image needs to perform the preprocessing operation, but also enrich the way of determining that the contour of the structural feature of the fetal ultrasound image needs to perform the preprocessing operation.
  • the categories of structural features that need to perform preprocessing operations may include one or more of the structural features of the cranial brain, the structural features of the abdominal circumference, the structural features of the cerebellum, and the structural features of the heart. Examples are not limited.
  • the categories of structural features that do not need to perform preprocessing operations may include one or more of lateral ventricle structural features, femur structural features, humerus structural features, and septum pellucidum structural features, which are not limited in this optional embodiment.
  • the contour information for example, resolution, sharpness, contrast, noise, etc.
  • the contour information for example, resolution, sharpness, contrast, noise, etc.
  • preprocessing conditions For example, when the contour of the structural feature of the fetal ultrasound image has noise and/or the contrast is lower than a determined contrast threshold (for example: 90%), it is determined that the structural feature does not meet the parameter measurement conditions, that is, the structural feature needs to be The contour of the preprocessing operation is performed.
  • the category of the structural feature of the fetal ultrasound image is the category of the structural feature that needs to execute the preprocessing condition and/or the contour of the structural feature does not satisfy the parameter measurement condition
  • the determined preprocessing conditions can improve the accuracy of determining that the structural features of the fetal ultrasound image need to be preprocessed, and can also enrich the way that the contours of the structural features of the fetal ultrasound image need to be preprocessed.
  • the profile of the required structural features is beneficial to further improve the measurement accuracy and reliability of the geometric parameters of the structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
  • the contour of each structural feature corresponds to multiple nodes. And, based on the determined preprocessing method, a preprocessing operation is performed on the contour of each structural feature to obtain the contour of the preprocessed structural feature, which may include:
  • the arc radius corresponding to the contour of each structural feature is greater than or equal to the determined arc radius threshold (for example: 5mm)
  • the way of connecting the two nodes connects all the target nodes corresponding to each structural feature in turn to obtain the preprocessed outline of the structural feature;
  • the contour of the structural feature when the contour of the structural feature has multiple arcs and/or the curvature of the contour is greater than or equal to the determined curvature threshold, the contour of the structural feature is segmented to perform a preprocessing operation. Specifically: when there are multiple arcs in the contour of the structural feature, the preprocessing operation will be performed on each of the multiple arcs of the structural feature; when the curvature of the contour of the structural feature is greater than or equal to the curvature threshold, the The contours of structural features are divided into multiple segments at equal or unequal intervals, and preprocessing operations are performed on each segment of the contour separately.
  • this optional embodiment can not only realize the preprocessing of structural features, but also improve the preprocessing efficiency and accuracy of structural features by selecting different preprocessing methods according to the size of the arc radius of the structural features of the fetal ultrasound image. , thereby improving the measurement accuracy of the geometric parameters of the structural features, and further improving the determination accuracy of the fetal growth and development situation corresponding to the fetal ultrasound image.
  • performing a preprocessing operation on the contour of each structural feature based on the determined preprocessing method, to obtain the contour of the preprocessed structural feature which may include:
  • a preprocessing operation is performed on the contour of each structural feature to obtain the contour of the preprocessed structural feature.
  • the B-spline curve fitting method may include an interpolation fitting method and/or an approximate fitting method, which is not limited in this optional embodiment.
  • this optional embodiment can not only achieve smooth fitting of the contour of the structural feature of the fetal ultrasound image, but also enrich the smooth fitting method of the contour of the structural feature.
  • implementing the automatic measurement method of fetal structural features described in FIG. 2 can automatically obtain the categories and contours of fetal structural features, and measure the contours of structural features according to the corresponding measurement methods, without manually obtaining the geometric parameters of fetal structural features. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results.
  • the noise in the contour of the structural feature can be filtered out, and a clear contour of the structural feature can be obtained, thereby helping to improve the efficiency and accuracy of obtaining geometric parameters of the structural feature.
  • FIG. 3 is a schematic structural diagram of an automatic measuring device for fetal structural features disclosed in an embodiment of the present invention. 3 can be applied to a parameter measurement server, wherein the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention.
  • the automatic measuring device for fetal structural features may include an analysis module 301, an acquisition module 302 and a measurement module 303, wherein:
  • the analysis module 301 inputs the acquired fetal ultrasound image into the determined feature detection model for analysis.
  • the acquisition module 302 is configured to acquire the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the category of at least one structural feature of the fetal ultrasound image and the at least one structural feature of the fetal ultrasound image. contour.
  • all structural features include at least one of cranial brain structural features, abdominal circumference structural features, and other structural features, and other structural features include cerebellar structural features, third ventricle structural features, and lateral ventricle structural features at least one of a feature, a septum pellucida feature, a humerus feature, and a femur feature.
  • the measurement module 303 is configured to determine a measurement mode that matches the contour of the structural feature according to the category of each structural feature, and perform a measurement operation on the contour of the structural feature according to the measurement mode of each structural feature, to obtain the contour of the structural feature. Geometric parameters.
  • the automatic measuring device implementing the fetal structural feature described in FIG. 3 can automatically obtain the category and contour of the fetal structural feature, and measure the contour of the structural feature according to the corresponding measurement method, without manually obtaining the geometrical parameters of the fetal structural feature. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results.
  • the measurement module 303 determines a measurement method that matches the outline of each structural feature according to the category of each structural feature, and determines the measurement method for the structure according to the measurement method of each structural feature.
  • the outline of the feature is measured, and the method to obtain the geometric parameters of the structural feature is as follows:
  • the geometric parameters corresponding to the cranial structure feature are determined based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature, and the contour of the cranial structure feature includes the outer contour and inner contour;
  • the perimeter corresponding to the contour of the abdominal circumference structural feature is obtained as the geometric parameter of the abdominal circumference structural feature
  • Geometric parameters of other structural features, and graph fitting methods include minimum circumscribed rectangle fitting methods and/or ellipse fitting methods.
  • the measurement module 303 fits the contour of each other structural feature based on the determined graphic fitting method, obtains the target node of the other structural feature, and determines the target node of each other structural feature based on the determined graph fitting method
  • the method of determining the geometric parameters of the other structural features is as follows:
  • the graph fitting mode is the minimum circumscribed rectangle fitting mode, based on the determined circumscribed rectangle fitting and the contours of other structural features, when a set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, Stop fitting the contours of other structural features, obtain the minimum circumscribed rectangle of the contours of other structural features, determine the scanning direction of the other structural features according to the category of the other structural features, and determine the scanning direction of the other structural features based on the scanning direction.
  • the contour scan operation is performed to obtain multiple scan line segments of the contour of the other structural feature, and the scan line segment with the largest distance value is selected from all the scan line segments as the geometric parameters of the other structural feature.
  • the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to represent a pair of short sides parallel to the smallest circumscribed rectangle.
  • the contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle, for example: scanning of the structural feature of lateral ventricle enlargement
  • the direction is parallel to the short side of the smallest circumscribed rectangle, and the scanning direction of the structural features of the left atrium is parallel to the long side of the smallest circumscribed rectangle.
  • the target intersection point is in the process of wirelessly approximating the contours of other structural features by the circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, stop fitting other structural features.
  • the outline of the structural feature can improve the accuracy of determining the scanning direction of the structural feature, thereby further improving the accuracy, reliability and efficiency of obtaining geometric parameters of the structural feature.
  • the measurement module 303 fits the contour of each other structural feature based on the determined graphic fitting method, obtains the target node of the other structural feature, and determines the target node of each other structural feature based on the determined graph fitting method
  • the method of determining the geometric parameters of the other structural features is as follows:
  • the graph fitting method is the ellipse fitting method
  • the contours of other structural features are fitted based on the determined ellipse fitting algorithm to obtain the contours of the fitted other structural features and the shape of the fitted contours of the other structural features. is oval;
  • the target scanning direction includes one of the first target scanning direction or the second target scanning direction, wherein the first target scanning direction is used to represent the direction from the contour parallel to the other structural features after fitting.
  • the short axis scans the contours of other structural features
  • the second target scanning direction is used to indicate that the contours of other structural features are scanned from the long axis parallel to the contours of the other structural features after fitting.
  • implementing the device described in FIG. 3 can also measure the geometric parameters of the corresponding structural features by measuring different geometric parameters according to different structural features.
  • the measuring module 303 determines a specific way of determining the geometric parameter corresponding to the cranial structure feature based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature for:
  • implementing the device described in FIG. 3 can also obtain the perimeter of the inner and outer contours of the structural features of the skull and the intersection of the mid-perpendicular line corresponding to the midline of the brain and the inner and outer contours of the structural features of the skull. Acquisition of head circumference and biparietal diameter length for structural features of the brain.
  • the apparatus may further include a preprocessing module 304, wherein:
  • the preprocessing module 304 is configured to, after the acquisition module 302 acquires the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, performs a preprocessing operation on the contour of each structural feature based on the determined preprocessing method to obtain a preprocessing operation.
  • implementing the device described in FIG. 4 can further perform a preprocessing operation on the contour of the structural feature of the fetal ultrasound image after acquiring the feature information of the fetal ultrasound image, such as a smoothing operation, which can convert the contour of the structural feature into a Noise filtering is used to obtain clear outlines of structural features, which is beneficial to improve the efficiency and accuracy of geometric parameter acquisition of structural features.
  • a preprocessing operation on the contour of the structural feature of the fetal ultrasound image after acquiring the feature information of the fetal ultrasound image, such as a smoothing operation, which can convert the contour of the structural feature into a Noise filtering is used to obtain clear outlines of structural features, which is beneficial to improve the efficiency and accuracy of geometric parameter acquisition of structural features.
  • the outline of each structural feature corresponds to multiple nodes; and, as shown in FIG. 4 , the preprocessing module 304 analyzes each structural feature based on the determined preprocessing method.
  • the preprocessing operation is performed on the contour of , and the method of obtaining the contour of the preprocessed structural feature is as follows:
  • each structural feature When the arc radius corresponding to the contour of each structural feature is greater than or equal to the determined arc radius threshold, select a preset number of target nodes from all nodes corresponding to the structural feature, and connect each adjacent two nodes. Connect all the target nodes corresponding to each structural feature in turn to obtain the preprocessed contour of the structural feature;
  • different preprocessing methods can be selected according to the size of the arc radius of the structural features of the fetal ultrasound image, which can not only realize the preprocessing of the structural features, but also can improve the preprocessing efficiency of the structural features and Therefore, the measurement accuracy of the geometric parameters of the structural features is improved, and the determination accuracy of the fetal growth and development situation corresponding to the fetal ultrasound image is further improved.
  • the apparatus may further include a judgment module 305 and a screening module 306, wherein:
  • the judgment module 305 is used to obtain the analysis result output by the feature detection model in the acquisition module 302 as the feature information of the fetal ultrasound image, and judge whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image. , when the judgment result is no, triggering the measurement module 303 to perform the above-mentioned operation of measuring the contour of the structural feature according to the measurement method of each structural feature to obtain the geometric parameters of the structural feature.
  • the screening module 306 is used to screen all the target structural features of the preprocessing condition from all the structural features when the judgment result of the judging module 305 is yes, and trigger the preprocessing module 304 to execute the above-mentioned preprocessing method based on the determined preprocessing method. Perform a preprocessing operation on the contour of each structural feature to obtain the contour of the preprocessed structural feature, and all structural features include all target structural features.
  • implementing the device described in FIG. 4 can further determine whether the structural features of the fetal ultrasound images satisfy the preprocessing conditions after acquiring the feature information of the fetal ultrasound images, and if the preprocessing conditions are not met, directly execute the structural features.
  • the contour measurement operation can improve the measurement efficiency of the geometric parameters of the structural features, thereby improving the determination efficiency of the growth and development of the fetus; Accurate and reliable contours of structural features are obtained, thereby improving the measurement accuracy and reliability of geometric parameters of structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
  • the manner in which the judgment module 305 judges whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image is specifically:
  • each structural feature it is determined whether the structural feature is the determined category of the structural feature that needs to perform a preprocessing operation, and when the judgment result is yes, it is determined that the structural feature satisfies the determined preprocessing condition; or,
  • each structural feature it is determined whether the contour of the structural feature satisfies the determined parameter measurement condition, and when the judgment result is no, it is determined that the structural feature satisfies the determined preprocessing condition.
  • implementing the device described in FIG. 4 can also be determined by determining that the category of the structural feature of the fetal ultrasound image is the category of the structural feature that needs to perform preprocessing conditions and/or the contour of the structural feature does not meet the parameter measurement conditions.
  • the structural features satisfy the determined preprocessing conditions, which can improve the accuracy of determining that the structural features of the fetal ultrasound image need to perform preprocessing operations, and can also enrich the way that the contours of the structural features of the fetal ultrasound images need to perform preprocessing operations, which is beneficial to Obtaining the structural features that meet the requirements is beneficial to further improve the measurement accuracy and reliability of the geometric parameters of the structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
  • FIG. 5 is another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention.
  • the automatic measurement device for fetal structural features described in FIG. 5 may be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention.
  • the automatic measuring device for fetal structural characteristics may include:
  • a memory 501 storing executable program code
  • processor 502 coupled to the memory 501;
  • an input interface 503 coupled with the processor 502 and an output interface 504;
  • the processor 502 invokes the executable program code stored in the memory 501 to execute some or all of the steps in the method for automatic measurement of fetal structural features described in Embodiment 1 or Embodiment 2.
  • An embodiment of the present invention discloses a computer-readable storage medium, which stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the automatic measurement method for fetal structural features described in the first or second embodiment some or all of the steps.
  • An embodiment of the present invention discloses a computer program product.
  • the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the description in the first embodiment or the second embodiment.
  • modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in One place, or it can be distributed over multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.
  • Read-Only Memory ROM
  • Random Access Memory Random Access Memory
  • PROM Programmable Read-only Memory
  • EPROM Erasable Programmable Read Only Memory
  • OTPROM One-time Programmable Read-Only Memory
  • EEPROM Electronically Erasable Programmable Read-Only Memory
  • CD-ROM Compact Disc Read -Only Memory

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Molecular Biology (AREA)
  • Data Mining & Analysis (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Artificial Intelligence (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Geometry (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)

Abstract

Disclosed in the present invention are an automatic measurement method and device for a fetal structural characteristic. The method comprises: inputting an acquired fetal ultrasound image into a determined characteristic detection model for analysis, and acquiring an analysis result output by the characteristic detection model as characteristic information of the fetal ultrasound image, the characteristic information of the fetal ultrasound image comprising the category of at least one structural characteristic of the fetal ultrasound image and the outline of the at least one structural characteristic; determining, according to the category of each structural characteristic, a measurement mode matching the outline of the structural characteristic; and performing a measurement operation on the outline of each structural characteristic according to the measurement mode of the structural characteristic to obtain a geometric parameter of the structural characteristic. Hence, a high-precision measurement result of the fetal structural characteristic can be rapidly acquired by implementing the present invention, thereby accurately determining the growth and development conditions of a fetus; inputting the fetal ultrasound image into the characteristic detection model can improve the efficiency of acquiring the measurement result of the fetal structural characteristic.

Description

一种胎儿结构特征的自动测量方法及装置A kind of automatic measurement method and device of fetal structural characteristics 技术领域technical field
本发明涉及图像技术领域，尤其涉及一种胎儿结构特征的自动测量方法及装置。The present invention relates to the field of image technology, and in particular, to an automatic measurement method and device for fetal structural features.
背景技术Background technique
随着社会的进步以及人们获取健康新生儿意识的加强，越来越多的孕妇按照产检计划定期前往医院进行产检来获知胎儿的生长发育情况。With the progress of society and the strengthening of people's awareness of obtaining healthy newborns, more and more pregnant women go to the hospital for regular obstetric examinations according to the obstetric examination plan to learn about the growth and development of the fetus.
目前，胎儿生长发育情况的获取方法为：检查者(例如：医护人员)通过移动轨迹球的方式手动确定胎儿头围或腹围的长轴和短轴，得到胎儿头围或腹围的椭圆形状，再通过椭圆周长公式计算胎儿头围或腹围周长，从而得到胎儿头围或腹围的测量结果，并根据该测量结果确定胎儿的生长发育情况。然而，实践发现，由于检查者的操作手法和熟练程度会影响对胎儿的结构特征的定位，这很容易导致获取到低精度的胎儿测量结果，从而无法准确确定胎儿的生长发育情况。At present, the acquisition method of fetal growth and development is as follows: the examiner (for example, a medical staff) manually determines the long axis and short axis of the fetal head circumference or abdominal circumference by moving the trackball, and obtains the oval shape of the fetal head circumference or abdominal circumference. , and then calculate the fetal head circumference or abdominal circumference through the ellipse circumference formula, so as to obtain the measurement result of the fetal head circumference or abdominal circumference, and determine the growth and development of the fetus according to the measurement result. However, it has been found in practice that since the operator's manipulation and proficiency will affect the positioning of the fetal structural features, it is easy to obtain low-precision fetal measurement results, and thus the fetal growth and development cannot be accurately determined.
发明内容SUMMARY OF THE INVENTION
本发明所要解决的技术问题在于，提供一种胎儿结构特征的自动测量方法及装置，能够获取到高精度的胎儿测量结果，从而准确确定胎儿的生长发育情况。The technical problem to be solved by the present invention is to provide an automatic measurement method and device for fetal structural characteristics, which can obtain high-precision fetal measurement results, so as to accurately determine the growth and development of the fetus.
为了解决上述技术问题，本发明第一方面公开了一种胎儿结构特征的自动测量方法，所述方法包括：In order to solve the above technical problems, the first aspect of the present invention discloses an automatic measurement method for fetal structural characteristics, the method comprising:
将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析，并获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息，所述胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓；The acquired fetal ultrasound image is input into the determined feature detection model for analysis, and the analysis result output by the feature detection model is obtained as the feature information of the fetal ultrasound image, and the feature information of the fetal ultrasound image includes the a category of at least one structural feature of the fetal ultrasound image and a contour of at least one of the structural features;
根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。According to the category of each of the structural features, a measurement method that matches the contour of the structural feature is determined, and a measurement operation is performed on the contour of the structural feature according to the measurement method of each of the structural features to obtain the geometric parameters of the structural feature. .
作为一种可选的实施方式，在本发明第一方面中，所述获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，所述方法还包括：As an optional implementation manner, in the first aspect of the present invention, after obtaining the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, the method further includes:
基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，并触发执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作，每个所述结构特征的轮廓均为预处理后的结构特征的轮廓，所述预处理操作包括平滑处理操作。Perform a preprocessing operation on the contour of each of the structural features based on the determined preprocessing method, obtain the contour of the preprocessed structural feature, and trigger the execution of the measurement method according to each structural feature for the structure A measurement operation is performed on the contour of the feature to obtain the geometric parameters of the structural feature. The contour of each structural feature is the contour of the structural feature after preprocessing, and the preprocessing operation includes a smoothing operation.
作为一种可选的实施方式，在本发明第一方面中，所述获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，所述方法还包 括：As an optional embodiment, in the first aspect of the present invention, after the described acquisition of the analysis result output by the feature detection model, as the feature information of the fetal ultrasound image, the method further includes:
根据所述胎儿超声图像的特征信息判断每个所述结构特征的轮廓是否满足确定出的预处理条件；Judging whether the contour of each of the structural features satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image;
当判断结果为否时，触发执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作；When the judgment result is no, trigger the execution of the described operation of performing the measurement operation on the outline of the structural feature according to the measurement method of each of the structural features to obtain the geometric parameters of the structural feature;
当判断结果为是时，从所有所述结构特征中筛选满足所述预处理条件的所有目标结构特征，并触发执行所述的基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的操作，所有所述结构特征包括所有所述目标结构特征。When the judgment result is yes, screen all the target structural features that satisfy the preprocessing conditions from all the structural features, and trigger the execution of the outline of each structural feature based on the determined preprocessing method. The preprocessing operation is an operation of obtaining the contours of the preprocessed structural features, where all the structural features include all the target structural features.
作为一种可选的实施方式，在本发明第一方面中，每个所述结构特征的轮廓均对应多个节点；As an optional implementation manner, in the first aspect of the present invention, the outline of each of the structural features corresponds to a plurality of nodes;
以及，所述基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，包括：And, performing a preprocessing operation on the contour of each of the structural features based on the determined preprocessing method, to obtain the contours of the preprocessed structural features, including:
获取每个所述结构特征的轮廓对应的圆弧半径；Obtain the arc radius corresponding to the contour of each of the structural features;
当每个所述结构特征的轮廓对应的圆弧半径大于等于确定出的圆弧半径阈值时，从该结构特征对应的所有所述节点中选取预设数量的目标节点，并按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述目标节点依次连接起来，得到预处理后的该结构特征的轮廓；When the arc radius corresponding to the outline of each of the structural features is greater than or equal to the determined arc radius threshold, select a preset number of target nodes from all the nodes corresponding to the structural feature, and select a preset number of target nodes according to each adjacent two All the target nodes corresponding to each of the structural features are sequentially connected in a manner of connecting each node to obtain the preprocessed outline of the structural feature;
当每个所述结构特征的轮廓对应的圆弧半径不大于等于确定出的圆弧半径阈值时，按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述节点依次连接起来，得到预处理后的该结构特征的轮廓。When the arc radius corresponding to the outline of each structural feature is not greater than or equal to the determined arc radius threshold, connect all the nodes corresponding to each structural feature in a manner of connecting every two adjacent nodes They are connected in sequence to obtain the preprocessed outline of the structural feature.
作为一种可选的实施方式，在本发明第一方面中，所有所述结构特征包括颅脑结构特征、腹围结构特征以及其他结构特征中的至少一种，所述其他结构特征包括小脑结构特征、第三脑室结构特征、侧脑室结构特征、透明隔结构特征、肱骨结构特征以及股骨结构特征中的至少一种。As an optional embodiment, in the first aspect of the present invention, all the structural features include at least one of the structural features of the cranial brain, the structural features of the abdominal circumference, and other structural features, and the other structural features include the structure of the cerebellum At least one of features, structural features of the third ventricle, structural features of the lateral ventricle, structural features of the septum pellucidum, structural features of the humerus, and structural features of the femur.
作为一种可选的实施方式，在本发明第一方面中，所述根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数，包括：As an optional implementation manner, in the first aspect of the present invention, the measurement method that matches the outline of each structural feature is determined according to the category of each structural feature, and The measurement method performs a measurement operation on the outline of the structural feature to obtain the geometric parameters of the structural feature, including:
当所述结构特征为所述颅脑结构特征时，基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数，所述颅脑结构特征的轮廓包括所述外轮廓和所述内轮廓；When the structural feature is the cranial structural feature, the geometric parameters corresponding to the cranial structural feature are determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature. The contour of the brain structure feature includes the outer contour and the inner contour;
当所述结构特征为所述腹围结构特征时，获取所述腹围结构特征的轮廓对应的周长，作为所述腹围结构特征的几何参数；When the structural feature is the abdominal circumference structural feature, acquiring the perimeter corresponding to the contour of the abdominal circumference structural feature as a geometric parameter of the abdominal circumference structural feature;
当所述结构特征为所述其他结构特征时，基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，所述图形拟合方式包括最小外接矩形拟合方式和/或椭圆拟合方式。When the structural feature is the other structural feature, fit the contour of each other structural feature based on the determined graph fitting method, obtain the target node of the other structural feature, and based on each of the other structural features The target node of the feature determines the geometric parameters of the other structural feature, and the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
作为一种可选的实施方式，在本发明第一方面中，所述基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，包括：As an optional implementation manner, in the first aspect of the present invention, the contour of each of the other structural features is fitted based on the determined graph fitting method, to obtain the target node of the other structural feature, and based on For each target node of the other structural feature, determine the geometric parameters of the other structural feature, including:
当所述图形拟合方式为所述最小外接矩形拟合方式时，基于确定出的外接矩形拟合所述其他结构特征的轮廓，当所述外接矩形的其中一组对边均与所述其他结构特征的轮廓存在目标交点时，停止拟合所述其他结构特征的轮廓，得到所述其他结构特征的轮廓的最小外接矩形，并根据所述其他结构特征的类别确定所述其他结构特征的扫查方向；When the graph fitting mode is the minimum circumscribed rectangle fitting mode, the contours of the other structural features are fitted based on the determined circumscribed rectangle. When the contour of the structural feature has a target intersection, stop fitting the contour of the other structural feature, obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the sweep of the other structural feature according to the category of the other structural feature. check the direction;
基于所述扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条扫查线段，并从所有所述扫查线段中选择距离值最大的扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the scanning direction, obtain multiple scanning line segments of the contours of the other structural features, and select the scanning line segment with the largest distance value from all the scanning line segments Line segments, as geometric parameters of the other structural features;
其中，所述扫查方向包括第一扫查方向或第二扫查方向中的其中一种，其中，所述第一扫查方向用于表示从平行于所述最小外接矩形的短边对所述其他结构特征的轮廓进行扫查，所述第二扫查方向用于表示从平行于所述最小外接矩形的长边对所述其他结构特征的轮廓进行扫查。Wherein, the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to indicate that the scanning direction is parallel to the short side of the minimum circumscribed rectangle. The contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle.
作为一种可选的实施方式，在本发明第一方面中，所述基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，包括：As an optional implementation manner, in the first aspect of the present invention, the contour of each of the other structural features is fitted based on the determined graph fitting method, to obtain the target node of the other structural feature, and based on For each target node of the other structural feature, determine the geometric parameters of the other structural feature, including:
当所述图形拟合方式为所述椭圆拟合方式时，基于确定出的椭圆拟合算法拟合所述其他结构特征的轮廓，得到拟合后的所述其他结构特征的轮廓，拟合后的所述其他结构特征的轮廓的形状为椭圆形；When the graph fitting method is the ellipse fitting method, the contours of the other structural features are fitted based on the determined ellipse fitting algorithm, and the fitted contours of the other structural features are obtained. The shape of the outline of the other structural features is an ellipse;
根据所述其他结构特征的类别确定所述其他结构特征的目标扫查方向；Determine the target scanning direction of the other structural features according to the categories of the other structural features;
基于所述目标扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条目标扫查线段，并从所有所述目标扫查线段中选择距离值最大的目标扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the target scanning direction, obtain multiple target scanning line segments of the contours of the other structural features, and select the largest distance value from all the target scanning line segments The target scanning line segment is used as the geometric parameter of the other structural features;
其中，所述目标扫查方向包括第一目标扫查方向或第二目标扫查方向中的其中一种，其中，所述第一目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的短轴对所述其他结构特征的轮廓进行扫查，所述第二目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的长轴对所述其他结构特征的轮廓进行扫查。Wherein, the target scanning direction includes one of a first target scanning direction or a second target scanning direction, wherein the first target scanning direction is used to represent a direction from parallel to the fitted The short axis of the contour of the other structural feature scans the contour of the other structural feature, and the second target scanning direction is used to indicate that the long axis of the contour of the other structural feature is parallel to the fitting. Scan the contours of other structural features described above.
作为一种可选的实施方式，在本发明第一方面中，所述基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数，包括：As an optional implementation manner, in the first aspect of the present invention, the geometry corresponding to the cranial structural feature is determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature parameters, including:
获取所述颅脑结构特征的内轮廓的第一周长与所述颅脑结构特征的外轮廓的第二周长，并基于所述第一周长和所述第二周长，确定所述颅脑结构特征对应的头围几何参数；Obtain the first circumference of the inner contour of the cranial structure feature and the second circumference of the outer contour of the skull structure feature, and determine the first circumference based on the first circumference and the second circumference The geometric parameters of the head circumference corresponding to the structural features of the brain;
确定所述颅脑结构特征的脑中线对应的中垂线与所述颅脑结构特征的外轮廓的第一交点以及所述中垂线与所述颅脑结构特征的内轮廓的第二交点，并基于所述第一交点与所述第二交点，确定所述颅脑结构特征对应的双顶径几何参数。determining the first intersection point of the midline corresponding to the midline of the cranial structure feature and the outer contour of the craniocerebral structure feature and the second intersection point of the perpendicular line and the inner contour of the craniocerebral structure feature, And based on the first intersection point and the second intersection point, the biparietal diameter geometric parameter corresponding to the cranial brain structural feature is determined.
本发明第二方面公开了一种胎儿结构特征的自动测量装置，所述装置包括：A second aspect of the present invention discloses an automatic measuring device for fetal structural features, the device comprising:
分析模块，将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析；An analysis module, which inputs the acquired fetal ultrasound image into the determined feature detection model for analysis;
获取模块，用于获取所述特征检测模型输出的分析结果，作为所述胎儿超 声图像的特征信息，所述胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓；The acquisition module is configured to acquire the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the category of at least one structural feature of the fetal ultrasound image and at least one of the features of the fetal ultrasound image. outline of structural features;
测量模块，用于根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。The measurement module is configured to determine a measurement mode that matches the contour of the structural feature according to the category of each of the structural features, and perform a measurement operation on the contour of the structural feature according to the measurement mode of each of the structural features to obtain the Geometric parameters of structural features.
作为一种可选的实施方式，在本发明第二方面中，所述装置还包括：As an optional implementation manner, in the second aspect of the present invention, the device further includes:
预处理模块，用于在所述获取模块获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，并触发所述测量模块执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作，每个所述结构特征的轮廓均为预处理后的结构特征的轮廓，所述预处理操作包括预处理操作。The preprocessing module is configured to, after the acquisition module acquires the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, executes on the contour of each of the structural features based on the determined preprocessing method Preprocessing operation, obtaining the contour of the preprocessed structural feature, and triggering the measurement module to perform the measurement operation on the contour of the structural feature according to the measurement method of each of the structural features, to obtain the geometry of the structural feature In the operation of parameters, the contour of each structural feature is the contour of the preprocessed structural feature, and the preprocessing operation includes a preprocessing operation.
作为一种可选的实施方式，在本发明第二方面中，所述装置还包括：As an optional implementation manner, in the second aspect of the present invention, the device further includes:
判断模块，用于在所述获取模块获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，根据所述胎儿超声图像的特征信息判断每个所述结构特征的轮廓是否满足确定出的预处理条件，当判断结果为否时，触发所述测量模块执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作；The judgment module is used for, after the acquisition module obtains the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, judges the contour of each of the structural features according to the feature information of the fetal ultrasound image Whether the determined preprocessing condition is met, when the judgment result is no, trigger the measurement module to perform the measurement operation for the contour of the structural feature according to the measurement method of each of the structural features, and obtain the profile of the structural feature. manipulation of geometric parameters;
筛选模块，用于当所述判断模块的判断结果为是时，从所有所述结构特征中筛选所述预处理条件的所有目标结构特征，并触发所述预处理模块执行所述的基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的操作，所有所述结构特征包括所有所述目标结构特征。The screening module is used for screening all the target structural features of the preprocessing condition from all the structural features when the judgment result of the judgment module is yes, and triggering the preprocessing module to perform the The preprocessing method is an operation of performing a preprocessing operation on the contour of each of the structural features to obtain the contours of the preprocessed structural features, and all the structural features include all the target structural features.
作为一种可选的实施方式，在本发明第二方面中，每个所述结构特征的轮廓均对应多个节点；As an optional implementation manner, in the second aspect of the present invention, the outline of each of the structural features corresponds to a plurality of nodes;
以及，所述预处理模块基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的方式具体为：And, the preprocessing module performs a preprocessing operation on the contour of each structural feature based on the determined preprocessing method, and the method for obtaining the contour of the preprocessed structural feature is specifically:
获取每个所述结构特征的轮廓对应的圆弧半径；Obtain the arc radius corresponding to the contour of each of the structural features;
当每个所述结构特征的轮廓对应的圆弧半径大于等于确定出的圆弧半径阈值时，从该结构特征对应的所有所述节点中选取预设数量的目标节点，并按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述目标节点依次连接起来，得到预处理后的该结构特征的轮廓；When the arc radius corresponding to the outline of each of the structural features is greater than or equal to the determined arc radius threshold, select a preset number of target nodes from all the nodes corresponding to the structural feature, and select a preset number of target nodes according to each adjacent two All the target nodes corresponding to each of the structural features are sequentially connected in a manner of connecting each node to obtain the preprocessed outline of the structural feature;
当每个所述结构特征的轮廓对应的圆弧半径不大于等于确定出的圆弧半径阈值时，按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述节点依次连接起来，得到预处理后的该结构特征的轮廓。When the arc radius corresponding to the outline of each structural feature is not greater than or equal to the determined arc radius threshold, connect all the nodes corresponding to each structural feature in a manner of connecting every two adjacent nodes They are connected in sequence to obtain the preprocessed outline of the structural feature.
作为一种可选的实施方式，在本发明第二方面中，所有所述结构特征包括颅脑结构特征、腹围结构特征以及其他结构特征中的至少一种，所述其他结构特征包括小脑结构特征、第三脑室结构特征、侧脑室结构特征、透明隔结构特征、肱骨结构特征以及股骨结构特征中的至少一种。As an optional embodiment, in the second aspect of the present invention, all the structural features include at least one of the structural features of the cranial brain, the structural features of the abdominal circumference, and other structural features, and the other structural features include the structure of the cerebellum At least one of features, structural features of the third ventricle, structural features of the lateral ventricle, structural features of the septum pellucidum, structural features of the humerus, and structural features of the femur.
作为一种可选的实施方式，在本发明第二方面中，所述测量模块根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征 的几何参数的方式具体为：As an optional implementation manner, in the second aspect of the present invention, the measurement module determines, according to the category of each of the structural features, a measurement method that matches the outline of the structural feature, and according to each of the structural features The measurement method of the feature performs a measurement operation on the outline of the structural feature, and the method for obtaining the geometric parameters of the structural feature is as follows:
当所述结构特征为所述颅脑结构特征时，基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数，所述颅脑结构特征的轮廓包括所述外轮廓和所述内轮廓；When the structural feature is the cranial structural feature, the geometric parameters corresponding to the cranial structural feature are determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature. The contour of the brain structure feature includes the outer contour and the inner contour;
当所述结构特征为所述腹围结构特征时，获取所述腹围结构特征的轮廓对应的周长，作为所述腹围结构特征的几何参数；When the structural feature is the abdominal circumference structural feature, acquiring the perimeter corresponding to the contour of the abdominal circumference structural feature as a geometric parameter of the abdominal circumference structural feature;
当所述结构特征为所述其他结构特征时，基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，所述图形拟合方式包括最小外接矩形拟合方式和/或椭圆拟合方式。When the structural feature is the other structural feature, fit the contour of each other structural feature based on the determined graph fitting method, obtain the target node of the other structural feature, and based on each of the other structural features The target node of the feature determines the geometric parameters of the other structural feature, and the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
作为一种可选的实施方式，在本发明第二方面中，所述测量模块基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数的方式具体为：As an optional implementation manner, in the second aspect of the present invention, the measurement module fits the contour of each of the other structural features based on the determined graphic fitting method, and obtains the target node of the other structural feature, And based on the target node of each of the other structural features, the specific way of determining the geometric parameters of the other structural features is:
当所述图形拟合方式为所述最小外接矩形拟合方式时，基于确定出的外接矩形拟合所述其他结构特征的轮廓，当所述外接矩形的其中一组对边均与所述其他结构特征的轮廓存在目标交点时，停止拟合所述其他结构特征的轮廓，得到所述其他结构特征的轮廓的最小外接矩形，并根据所述其他结构特征的类别确定所述其他结构特征的扫查方向；When the graph fitting mode is the minimum circumscribed rectangle fitting mode, the contours of the other structural features are fitted based on the determined circumscribed rectangle. When the contour of the structural feature has a target intersection, stop fitting the contour of the other structural feature, obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the sweep of the other structural feature according to the category of the other structural feature. check the direction;
基于所述扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条扫查线段，并从所有所述扫查线段中选择距离值最大的扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the scanning direction, obtain multiple scanning line segments of the contours of the other structural features, and select the scanning line segment with the largest distance value from all the scanning line segments Line segments, as geometric parameters of the other structural features;
其中，所述扫查方向包括第一扫查方向或第二扫查方向中的其中一种，其中，所述第一扫查方向用于表示从平行于所述最小外接矩形的短边对所述其他结构特征的轮廓进行扫查，所述第二扫查方向用于表示从平行于所述最小外接矩形的长边对所述其他结构特征的轮廓进行扫查。Wherein, the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to indicate that the scanning direction is parallel to the short side of the minimum circumscribed rectangle. The contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle.
作为一种可选的实施方式，在本发明第二方面中，所述测量模块基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数的方式具体为：As an optional implementation manner, in the second aspect of the present invention, the measurement module fits the contour of each of the other structural features based on the determined graphic fitting method, and obtains the target node of the other structural feature, And based on the target node of each of the other structural features, the specific way of determining the geometric parameters of the other structural features is:
当所述图形拟合方式为所述椭圆拟合方式时，基于确定出的椭圆拟合算法拟合所述其他结构特征的轮廓，得到拟合后的所述其他结构特征的轮廓，拟合后的所述其他结构特征的轮廓的形状为椭圆形；When the graph fitting method is the ellipse fitting method, the contours of the other structural features are fitted based on the determined ellipse fitting algorithm, and the fitted contours of the other structural features are obtained. The shape of the outline of the other structural features is an ellipse;
根据所述其他结构特征的类别确定所述其他结构特征的目标扫查方向；Determine the target scanning direction of the other structural features according to the categories of the other structural features;
基于所述目标扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条目标扫查线段，并从所有所述目标扫查线段中选择距离值最大的目标扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the target scanning direction, obtain multiple target scanning line segments of the contours of the other structural features, and select the largest distance value from all the target scanning line segments The target scanning line segment is used as the geometric parameter of the other structural features;
其中，所述目标扫查方向包括第一目标扫查方向或第二目标扫查方向中的其中一种，其中，所述第一目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的短轴对所述其他结构特征的轮廓进行扫查，所述第二目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的长轴对所述其他结构特征的轮廓进行扫查。Wherein, the target scanning direction includes one of a first target scanning direction or a second target scanning direction, wherein the first target scanning direction is used to represent a direction from parallel to the fitted The short axis of the contour of the other structural feature scans the contour of the other structural feature, and the second target scanning direction is used to indicate that the long axis of the contour of the other structural feature is parallel to the fitting. Scan the contours of other structural features described above.
作为一种可选的实施方式，在本发明第二方面中，所述测量模块基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数的方式具体为：As an optional implementation manner, in the second aspect of the present invention, the measurement module determines the corresponding cranial structure feature based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature The way of the geometric parameters is as follows:
获取所述颅脑结构特征的内轮廓的第一周长与所述颅脑结构特征的外轮廓的第二周长，并基于所述第一周长和所述第二周长，确定所述颅脑结构特征对应的头围几何参数；Obtain the first circumference of the inner contour of the cranial structure feature and the second circumference of the outer contour of the skull structure feature, and determine the first circumference based on the first circumference and the second circumference The geometric parameters of the head circumference corresponding to the structural features of the brain;
确定所述颅脑结构特征的脑中线对应的中垂线与所述颅脑结构特征的外轮廓的第一交点以及所述中垂线与所述颅脑结构特征的内轮廓的第二交点，并基于所述第一交点与所述第二交点，确定所述颅脑结构特征对应的双顶径几何参数。determining the first intersection point of the midline corresponding to the midline of the cranial structure feature and the outer contour of the craniocerebral structure feature and the second intersection point of the perpendicular line and the inner contour of the craniocerebral structure feature, And based on the first intersection point and the second intersection point, the biparietal diameter geometric parameter corresponding to the cranial brain structural feature is determined.
本发明第三方面公开了另一种胎儿结构特征的自动测量装置，所述装置包括：A third aspect of the present invention discloses another automatic measuring device for fetal structural features, the device comprising:
存储有可执行程序代码的存储器；a memory in which executable program code is stored;
与所述存储器耦合的处理器；a processor coupled to the memory;
所述处理器调用所述存储器中存储的所述可执行程序代码，执行本发明第一方面公开的胎儿结构特征的自动测量方法。The processor invokes the executable program code stored in the memory to execute the automatic measurement method for fetal structural features disclosed in the first aspect of the present invention.
本发明第四方面公开了一种计算机存储介质，所述计算机存储介质存储有计算机指令，所述计算机指令被调用时，用于执行本发明第一方面公开的胎儿结构特征的自动测量方法。A fourth aspect of the present invention discloses a computer storage medium, the computer storage medium stores computer instructions, and when the computer instructions are invoked, is used to execute the automatic measurement method for fetal structural features disclosed in the first aspect of the present invention.
与现有技术相比，本发明实施例具有以下有益效果：Compared with the prior art, the embodiments of the present invention have the following beneficial effects:
本发明实施例中，将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析，并获取该特征检测模型输出的分析结果，作为胎儿超声图像的特征信息，该胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓；根据每个结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。可见，实施本发明通过自动获取胎儿的结构特征的类别和轮廓，并根据对应的测量方式测量结构特征的轮廓，无需通过手动获取胎儿结构特征的几何参数，能够快速获取到高精度的胎儿结构特征的测量结果，从而准确确定胎儿的生长发育情况；以及通过将胎儿超声图像输入特征检测模型，能够提高胎儿的结构特征的测量结果的获取效率。In the embodiment of the present invention, the acquired fetal ultrasound image is input into the determined feature detection model for analysis, and the analysis result output by the feature detection model is obtained as the feature information of the fetal ultrasound image, the feature information of the fetal ultrasound image Including the category of at least one structural feature of the fetal ultrasound image and the contour of at least one structural feature; determining a measurement mode matching the contour of the structural feature according to the category of each structural feature, and according to the measurement mode of each structural feature A measurement operation is performed on the contour of the structural feature to obtain the geometrical parameters of the structural feature. It can be seen that the implementation of the present invention can quickly obtain high-precision fetal structural features by automatically acquiring the categories and contours of the structural features of the fetus, and measuring the contours of the structural features according to the corresponding measurement methods, without manually obtaining the geometric parameters of the fetal structural features. The measurement results of the fetus can be accurately determined, so as to accurately determine the growth and development of the fetus; and by inputting the fetal ultrasound image into the feature detection model, the acquisition efficiency of the measurement results of the fetal structural features can be improved.
附图说明Description of drawings
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to illustrate the technical solutions in the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative effort.
图1是本发明实施例公开的一种胎儿结构特征的自动测量方法的流程示意图；1 is a schematic flowchart of an automatic measurement method for fetal structural features disclosed in an embodiment of the present invention;
图2是本发明实施例公开的另一种胎儿结构特征的自动测量方法的流程示意图；2 is a schematic flowchart of another automatic measurement method for fetal structural features disclosed in an embodiment of the present invention;
图3是本发明实施例公开的一种胎儿结构特征的自动测量装置的结构示意 图；3 is a schematic structural diagram of an automatic measuring device for fetal structural features disclosed in an embodiment of the present invention;
图4是本发明实施例公开的另一种胎儿结构特征的自动测量装置的结构示意图；4 is a schematic structural diagram of another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention;
图5是本发明实施例公开的又一种胎儿结构特征的自动测量装置的结构示意图。FIG. 5 is a schematic structural diagram of another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention.
具体实施方式detailed description
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make those skilled in the art better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only These are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别不同对象，而不是用于描述特定顺序。此外，术语“包括”和“具有”以及它们任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、装置、产品或设备没有限定于已列出的步骤或单元，而是可选地还包括没有列出的步骤或单元，或可选地还包括对于这些过程、方法、产品或设备固有的其他步骤或单元。The terms "first", "second" and the like in the description and claims of the present invention and the above drawings are used to distinguish different objects, rather than to describe a specific order. Furthermore, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes unlisted steps or units, or optionally also includes For other steps or units inherent to these processes, methods, products or devices.
在本文中提及“实施例”意味着，结合实施例描述的特定特征、结构或特性可以包含在本发明的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例，也不是与其它实施例互斥的独立的或备选的实施例。本领域技术人员显式地和隐式地理解的是，本文所描述的实施例可以与其它实施例相结合。Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
本发明公开了一种胎儿结构特征的自动测量方法及装置，能够通过自动获取胎儿的结构特征的类别和轮廓，并根据对应的测量方式测量结构特征的轮廓，无需通过手动获取胎儿结构特征的几何参数，能够快速获取到高精度的胎儿结构特征的测量结果，从而准确确定胎儿的生长发育情况；以及通过将胎儿超声图像输入特征检测模型，能够提高胎儿的结构特征的测量结果的获取效率。以下分别进行详细说明。The invention discloses an automatic measuring method and device for fetal structural features, which can automatically obtain the category and contour of fetal structural features, and measure the contours of structural features according to corresponding measurement methods, without manually obtaining the geometrical features of fetal structural features. parameters, can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, the acquisition efficiency of fetal structural features can be improved. Each of them will be described in detail below.
实施例一Example 1
请参阅图1，图1是本发明实施例公开的一种胎儿结构特征的自动测量方法的流程示意图。其中，图1所描述的胎儿结构特征的自动测量方法可以应用于参数测量服务器中，其中，该参数测量服务器可以包括本地参数测量服务器或云参数测量服务器，本发明实施例不做限定。如图1所示，该胎儿结构特征的自动测量方法可以包括以下操作：Please refer to FIG. 1 . FIG. 1 is a schematic flowchart of an automatic measurement method for fetal structural features disclosed in an embodiment of the present invention. 1 can be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention. As shown in FIG. 1, the automatic measurement method of fetal structural features may include the following operations:
101、将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析，并获取特征检测模型输出的分析结果，作为该胎儿超声图像的特征信息，该胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓。101. Input the acquired fetal ultrasound image into the determined feature detection model for analysis, and obtain the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, which includes the fetal ultrasound image. The category of at least one structural feature of the image and the contour of at least one structural feature.
本发明实施例中，胎儿超声图像可以为单帧图片，也可以为动态图像。其中，当胎儿超声图像为单帧图片时，可以按照预先确定出的帧率将胎儿超声图像连续输入特征检测模型中进行分析，这样通过将连续多帧的胎儿超声图像输 入特征检测模型进行分析，有利于对同一结构特征进行多次分析，从而有利于对同一结构特征的轮廓进行多次测量，从而提高胎儿超声图像的结构特征的测量准确性，进而提高胎儿的生长发育情况的确定准确性；还可以对胎儿超声图像的更多结构特征的轮廓执行测量操作，从而有利于根据不同结构特征的几何参数确定胎儿超声图像对应的胎儿的生长发育情况。又进一步的，预先确定出的帧率与所需测量的胎儿超声图像的结构特征有关，即根据所需测量的胎儿超声图像的结构特征来选择帧率，例如：若需要测量的是腹围结构特征，则帧率可以为30帧/秒；若需要测量的是四腔心结构特征，则帧率可以为60帧/秒。这样根据所需测量的胎儿超声图像的结构特征选择对应的帧率，有利于提高所需胎儿超声图像的结构特征的几何参数的测量效率以及准确性。当胎儿超声图像为动态图像，特征检测模型在分析胎儿超声图像的特征信息之前，可以先将胎儿超声图像分为多帧胎儿超声图片。这样通过对动态的胎儿超声图像执行帧分割操作，能够对胎儿超声图像的结构特征进行细分，从而有利于提高胎儿超声图像的结构特征的几何参数的测量精准性。In this embodiment of the present invention, the fetal ultrasound image may be a single frame picture or a dynamic image. Wherein, when the fetal ultrasound image is a single-frame picture, the fetal ultrasound image can be continuously input into the feature detection model for analysis according to the predetermined frame rate, so that by inputting the continuous multiple frames of fetal ultrasound images into the feature detection model for analysis, It is beneficial to carry out multiple analysis of the same structural feature, thereby facilitating multiple measurement of the contour of the same structural feature, thereby improving the measurement accuracy of the structural feature of the fetal ultrasound image, thereby improving the accuracy of determining the growth and development of the fetus; The measurement operation can also be performed on the contours of more structural features of the fetal ultrasound image, thereby facilitating determining the growth and development of the fetus corresponding to the fetal ultrasound image according to geometric parameters of different structural features. Still further, the predetermined frame rate is related to the structural characteristics of the fetal ultrasound image to be measured, that is, the frame rate is selected according to the structural characteristics of the fetal ultrasound image to be measured, for example, if the abdominal circumference structure needs to be measured. feature, the frame rate can be 30 frames/second; if the four-chamber structural feature needs to be measured, the frame rate can be 60 frames/second. In this way, the corresponding frame rate is selected according to the structural features of the fetal ultrasound image to be measured, which is beneficial to improve the measurement efficiency and accuracy of the geometric parameters of the structural features of the fetal ultrasound image. When the fetal ultrasound image is a dynamic image, the feature detection model may first divide the fetal ultrasound image into multiple frames of fetal ultrasound pictures before analyzing the feature information of the fetal ultrasound image. In this way, by performing the frame segmentation operation on the dynamic fetal ultrasound image, the structural features of the fetal ultrasound image can be subdivided, thereby helping to improve the measurement accuracy of the geometric parameters of the structural features of the fetal ultrasound image.
本发明实施例中，可选的，胎儿超声图像的特征信息还可以包括该胎儿超声图像的结构特征的图形坐标、尺寸、位置概率中的至少一种，本发明实施例不做限定。其中，图形坐标可以包括多边形坐标或椭圆形坐标，其中，多边形坐标可以包括奇数多边形坐标或偶数多边形坐标，例如：五角形坐标、长方形坐标，多边形坐标的选取取决于结构特征的形状，这样能够提高结构特征的坐标获取准确性。又进一步的，胎儿超声图像的特征信息还包括该胎儿超声图像的结构特征的多边形轮廓信息，例如：多边形轮廓坐标，这样胎儿超声图像的特征信息包括的内容越多越有利于提高胎儿的生长发育情况的确定准确性。In the embodiment of the present invention, optionally, the feature information of the fetal ultrasound image may further include at least one of graphic coordinates, size, and position probability of the structural feature of the fetal ultrasound image, which is not limited in the embodiment of the present invention. Wherein, the graphic coordinates may include polygon coordinates or elliptical coordinates, wherein the polygon coordinates may include odd polygon coordinates or even polygon coordinates, such as: pentagonal coordinates, rectangular coordinates, and the selection of polygon coordinates depends on the shape of the structural feature, which can improve the structure The coordinate acquisition accuracy of the feature. Still further, the feature information of the fetal ultrasound image also includes polygonal contour information of the structural features of the fetal ultrasound image, such as polygonal contour coordinates, so that the more content the feature information of the fetal ultrasound image includes, the more conducive to improving the growth and development of the fetus. The accuracy of the determination of the situation.
本发明实施例中，可选的，每帧胎儿超声图像均存在唯一对应的帧序号，这样通过为每帧胎儿超声图像设定唯一的帧序号，能够在胎儿超声图像的结构特征的测量过程中，清楚区分每帧胎儿超声图像以及有利于对胎儿超声图像及其结构特征的相关信息(例如：几何参数)的进行管理。In this embodiment of the present invention, optionally, each frame of fetal ultrasound image has a unique corresponding frame sequence number, so that by setting a unique frame sequence number for each frame of fetal ultrasound image, it is possible to measure the structural features of the fetal ultrasound image during the measurement process of the fetal ultrasound image. , clearly distinguish each frame of fetal ultrasound images and facilitate the management of fetal ultrasound images and related information (eg, geometric parameters) of their structural features.
本发明实施例中，所有结构特征可以包括颅脑结构特征、腹围结构特征以及其他结构特征中的至少一种，其他结构特征包括小脑结构特征、第三脑室结构特征、侧脑室结构特征、透明隔结构特征、肱骨结构特征以及股骨结构特征中的至少一种，本发明实施例不做限定。In this embodiment of the present invention, all structural features may include at least one of cranial brain structural features, abdominal circumference structural features, and other structural features, and other structural features include cerebellar structural features, third ventricle structural features, lateral ventricle structural features, transparency At least one of the structural features of the septum, the structural features of the humerus, and the structural features of the femur is not limited in the embodiment of the present invention.
102、根据每个结构特征的类别确定与该结构特征的轮廓相匹配的测量方式。102. Determine, according to the category of each structural feature, a measurement mode that matches the contour of the structural feature.
本发明实施例中，每个结构特征的轮廓均存在与之对应的测量方式。需要说明的是，获取到的结构特无论是正常的结构特征，还是异常的结构特征，其对应的测量方式均相同。In the embodiment of the present invention, the contour of each structural feature has a corresponding measurement method. It should be noted that, whether the acquired structural features are normal structural features or abnormal structural features, the corresponding measurement methods are the same.
103、根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。103. Perform a measurement operation on the contour of the structural feature according to the measurement method of each structural feature, to obtain geometric parameters of the structural feature.
可见，实施图1所描述的胎儿结构特征的自动测量方法能够通过自动获取胎儿的结构特征的类别和轮廓，并根据对应的测量方式测量结构特征的轮廓，无需通过手动获取胎儿结构特征的几何参数，能够快速获取到高精度的胎儿结构特征的测量结果，从而准确确定胎儿的生长发育情况；以及通过将胎儿超声图像输入特征检测模型，能够提高胎儿的结构特征的测量结果的获取效率。It can be seen that the automatic measurement method of fetal structural features described in FIG. 1 can automatically obtain the category and contour of fetal structural features, and measure the contour of structural features according to the corresponding measurement method, without manually obtaining the geometric parameters of fetal structural features. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results.
在一个可选的实施例中，根据每个结构特征的类别确定与该结构特征的轮 廓相匹配的测量方式，并根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数，可以包括：In an optional embodiment, a measurement method that matches the contour of each structural feature is determined according to the category of each structural feature, and a measurement operation is performed on the contour of the structural feature according to the measurement method of each structural feature to obtain the Geometric parameters of structural features, which can include:
当上述结构特征为颅脑结构特征时，基于颅脑结构特征的内轮廓与颅脑结构特征的外轮廓，确定颅脑结构特征对应的几何参数，该颅脑结构特征的轮廓包括颅脑结构特征的外轮廓和颅脑结构特征的内轮廓；When the above-mentioned structural feature is a cranial structure feature, the geometric parameters corresponding to the cranial structure feature are determined based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature, and the outline of the cranial structure feature includes the cranial structure feature The outer contour of the brain and the inner contour of the structural features of the brain;
当上述结构特征为腹围结构特征时，获取腹围结构特征的轮廓对应的周长，作为腹围结构特征的几何参数；When the above-mentioned structural feature is the abdominal circumference structural feature, the perimeter corresponding to the contour of the abdominal circumference structural feature is obtained as the geometric parameter of the abdominal circumference structural feature;
当上述结构特征为其他结构特征时，基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数。When the above-mentioned structural features are other structural features, fit the contour of each other structural feature based on the determined graph fitting method to obtain the target node of the other structural feature, and determine the target node of the other structural feature based on the target node of each other structural feature. Geometric parameters of other structural features.
该可选的实施例中，图形拟合方式包括最小外接矩形拟合方式和/或椭圆拟合方式。In this optional embodiment, the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
作为一种可选的实施方式，当图形拟合方式为最小外接矩形拟合方式时，基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数，包括：基于确定出的外接矩形拟合与其他结构特征的轮廓，当外接矩形的其中一组对边均与其他结构特征的轮廓存在目标交点时，停止拟合其他结构特征的轮廓，得到其他结构特征的轮廓的最小外接矩形，并根据该其他结构特征的类别确定该其他结构特征的扫查方向，并基于该扫查方向对该其他结构特征的轮廓执行扫查操作，得到该其他结构特征的轮廓的多条扫查线段，并从所有扫查线段中选择距离值最大的扫查线段，作为该其他结构特征的几何参数。As an optional embodiment, when the graph fitting method is the minimum circumscribed rectangle fitting method, the contour of each other structural feature is fitted based on the determined graph fitting method, and the target node of the other structural feature is obtained, And based on the target node of each other structural feature, determine the geometric parameters of the other structural feature, including: fitting the contour of the other structural feature based on the determined circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle is consistent with other structural features. When the contour of the feature has a target intersection, stop fitting the contour of other structural features, obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the scanning direction of the other structural feature according to the category of the other structural feature, and based on the scanning Perform a scanning operation on the contour of the other structural feature to obtain multiple scanning line segments of the contour of the other structural feature, and select the scanning line segment with the largest distance value from all the scanning line segments as the scanning line segment of the other structural feature. Geometric parameters.
该可选的实施方式中，扫查方向包括第一扫查方向或第二扫查方向中的其中一种，其中，该第一扫查方向用于表示从平行于最小外接矩形的短边对该其他结构特征的轮廓进行扫查，该第二扫查方向用于表示从平行于最小外接矩形的长边对该其他结构特征的轮廓进行扫查，例如：侧脑室增宽结构特征的扫查方向为平行于最小外接矩形的短边，左心房结构特征的扫查方向为平行于最小外接矩形的长边。又可选的，该目标交点为在外接矩形无线逼近其他结构特征的轮廓的过程中，当外接矩形的其中一组对边一旦均与其他结构特征的轮廓存在目标交点时，就停止拟合其他结构特征的轮廓，这样能够提高结构特征的扫查方向的确定准确性，从而进一步提高结构特征的几何参数的获取准确性、可靠性以及效率。In this optional implementation manner, the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to represent a pair of short sides parallel to the smallest circumscribed rectangle. The contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle, for example: scanning of the structural feature of lateral ventricle enlargement The direction is parallel to the short side of the smallest circumscribed rectangle, and the scanning direction of the structural features of the left atrium is parallel to the long side of the smallest circumscribed rectangle. Alternatively, the target intersection point is in the process of wirelessly approximating the contours of other structural features by the circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, stop fitting other structural features. The outline of the structural feature can improve the accuracy of determining the scanning direction of the structural feature, thereby further improving the accuracy, reliability and efficiency of obtaining geometric parameters of the structural feature.
可见，该可选的实施方式通过最小外接矩形拟合结构特征，能够实现结构特征的几何参数的获取，以及不同类别的结构特征选择不同的扫查方向，能够提高结构特征的几何参数的获取准确性、可靠性以及效率。It can be seen that this optional embodiment can obtain the geometric parameters of the structural features by fitting the structural features through the minimum circumscribed rectangle, and select different scanning directions for different types of structural features, which can improve the accuracy of the geometric parameters of the structural features. performance, reliability and efficiency.
该可选的实施例中，当上述其他结构特征为小脑结构特征和/或第三脑室结构特征时，其他结构特征的几何参数的获取方式还可以为：获取其他结构特征的脑中线对应的中垂线，并获取该中垂线与其他结构特征的轮廓的两个交点，以及连接该两个交点所形成的线段，作为其他结构特征的几何参数。进一步的，当其他结构特征的几何参数的获取方式包括上述两种获取方式时，获取该两种获取方式所获得其他结构特征的几何参数的均值，作为其他结构特征的最终横径几何参数。举例来说，当小脑横径的获取方式包括上述两种获取方式时，获 取该两种获取方式所获得小脑横径的均值，作为小脑结构特征的最终横径，从而提高其他结构特征的几何参数的获取准确性以及可靠性。In this optional embodiment, when the above-mentioned other structural features are cerebellar structural features and/or structural features of the third ventricle, the geometric parameters of the other structural features may also be obtained by: The vertical line is obtained, and the two intersection points of the vertical line and the outline of other structural features, as well as the line segment formed by connecting the two intersection points, are obtained as the geometric parameters of other structural features. Further, when the acquisition methods of the geometric parameters of other structural features include the above two acquisition methods, the mean value of the geometric parameters of the other structural features obtained by the two acquisition methods is acquired as the final transverse diameter geometric parameters of the other structural features. For example, when the acquisition methods of the cerebellar transverse diameter include the above two acquisition methods, the average value of the cerebellar transverse diameters obtained by the two acquisition methods is obtained as the final transverse diameter of the cerebellum structural features, thereby improving the geometric parameters of other structural features. accuracy and reliability of acquisition.
该可选的实施例中，可选的，基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数的方式具体为：In this optional embodiment, optionally, the contour of each other structural feature is fitted based on the determined graph fitting method to obtain the target node of the other structural feature, and based on the target node of each other structural feature, The specific way of determining the geometric parameters of the other structural features is as follows:
当图形拟合方式为椭圆拟合方式时，基于确定出的椭圆拟合算法拟合其他结构特征的轮廓，得到拟合后的其他结构特征的轮廓，拟合后的其他结构特征的轮廓的形状为椭圆形；When the graph fitting method is the ellipse fitting method, the contours of other structural features are fitted based on the determined ellipse fitting algorithm to obtain the contours of the fitted other structural features and the shape of the fitted contours of the other structural features. is oval;
根据其他结构特征的类别确定其他结构特征的目标扫查方向；Determine the target scanning direction of other structural features according to the categories of other structural features;
基于目标扫查方向对其他结构特征的轮廓执行扫查操作，得到其他结构特征的轮廓的多条目标扫查线段，并从所有目标扫查线段中选择距离值最大的目标扫查线段，作为其他结构特征的几何参数；Perform the scanning operation on the contours of other structural features based on the scanning direction of the target, obtain multiple target scanning line segments of the contours of other structural features, and select the target scanning line segment with the largest distance value from all the target scanning line segments as the other scanning line segments. geometric parameters of structural features;
其中，目标扫查方向包括第一目标扫查方向或第二目标扫查方向中的其中一种，其中，第一目标扫查方向用于表示从平行于拟合后的其他结构特征的轮廓的短轴对其他结构特征的轮廓进行扫查，第二目标扫查方向用于表示从平行于拟合后的其他结构特征的轮廓的长轴对其他结构特征的轮廓进行扫查。例如：颅脑结构特征的扫查方向为平行于椭圆拟合后的颅脑结构特征的轮廓的短轴，并把距离值最大的扫查线段作为颅脑结构特征的横径。Wherein, the target scanning direction includes one of the first target scanning direction or the second target scanning direction, wherein the first target scanning direction is used to represent the direction from the contour parallel to the other structural features after fitting. The short axis scans the contours of other structural features, and the second target scanning direction is used to indicate that the contours of other structural features are scanned from the long axis parallel to the contours of the other structural features after fitting. For example, the scanning direction of the cranial structural feature is parallel to the short axis of the contour of the cranial structural feature after ellipse fitting, and the scanning line segment with the largest distance value is taken as the transverse diameter of the cranial structural feature.
该可选的实施例中，需要说明的是，针对同一其他结构特征，还可以基于最小外接矩形拟合方式和椭圆拟合方式共同确定其他结构特征的扫查方向，例如：小脑横径的扫查方向可以由最小外接矩形和颅骨椭圆光环结构共同确定。这样通过椭圆拟合方式和最小外接矩形拟合的方式共同确定结构特征的扫查方向，能够提高结构特征的扫查方向的确定准确性以及可靠性，进而提高结构特征的几何参数的确定准确性以及可靠性。In this optional embodiment, it should be noted that, for the same other structural feature, the scanning direction of other structural features can also be determined based on the minimum circumscribed rectangle fitting method and the ellipse fitting method, for example, the scanning direction of the cerebellar transverse diameter. The direction of the search can be determined by the minimum circumscribed rectangle and the elliptical halo structure of the skull. In this way, the scanning direction of the structural feature is determined jointly by the ellipse fitting method and the minimum circumscribed rectangle fitting method, which can improve the accuracy and reliability of the scanning direction of the structural feature, thereby improving the accuracy of determining the geometric parameters of the structural feature. and reliability.
可见，该可选的实施例通过根据不同结构特征，采取不同几何参数的测量方式测量对应结构特征的几何参数，不仅能够实现胎儿超声图像的结构特征的几何参数的获取，还能够提高胎儿超声图像的结构特征的几何参数的获取准确性以及效率；以及通过椭圆拟合方式和/或最小外接矩形拟合方式等方式获取结构特征的几何参数，能够丰富结构特征的几何参数获取方式，以及提高结构特征的几何参数获取可能性以及准确性。It can be seen that this optional embodiment measures the geometric parameters of the corresponding structural features by adopting different geometric parameter measurement methods according to different structural features, which can not only achieve the acquisition of the geometric parameters of the structural features of the fetal ultrasound image, but also can improve the fetal ultrasound image. Accuracy and efficiency of obtaining geometric parameters of structural features; and obtaining geometric parameters of structural features by means of ellipse fitting and/or minimum circumscribed rectangle fitting, which can enrich the way of obtaining geometric parameters of structural features and improve the structure of Possibility and accuracy of geometric parameter acquisition of features.
在另一个可选的实施例中，基于颅脑结构特征的内轮廓与颅脑结构特征的外轮廓，确定颅脑结构特征对应的几何参数，可以包括：In another optional embodiment, based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature, determining the geometric parameters corresponding to the cranial structure feature may include:
获取颅脑结构特征的内轮廓的第一周长与颅脑结构特征的外轮廓的第二周长，并基于该第一周长和该第二周长，确定颅脑结构特征对应的头围几何参数；Obtain the first circumference of the inner contour of the skull structure feature and the second circumference of the outer contour of the skull structure feature, and determine the head circumference corresponding to the skull structure feature based on the first circumference and the second circumference Geometric parameters;
确定颅脑结构特征的脑中线对应的中垂线与颅脑结构特征的外轮廓的第一交点以及中垂线与颅脑结构特征的内轮廓的第二交点，并基于该第一交点与该第二交点，确定颅脑结构特征对应的双顶径几何参数。Determine the first intersection point of the midline corresponding to the midline of the craniocerebral structural feature and the outer contour of the craniocerebral structural feature, and the second intersection of the perpendicular line and the inner contour of the craniocerebral structural feature, and based on the first intersection and the The second intersection point is to determine the geometric parameters of the biparietal diameter corresponding to the structural features of the skull.
该可选的实施例中，颅脑结构特征对应的头围几何参数的计算方式如下：In this optional embodiment, the calculation method of the head circumference geometric parameter corresponding to the cranial structure feature is as follows:
C＝(C 1+C 2)/2； C=(C 1 +C 2 )/2;
式中，C为颅脑结构特征的头围几何参数，即头围周长；C 1为颅脑结构特征的外轮廓的第二周长；C 2颅脑结构特征的内轮廓的第一周长。 In the formula, C is the head circumference geometric parameter of the cranial structure feature, namely the head circumference; C 1 is the second circumference of the outer contour of the cranial structure feature; C 2 is the first circumference of the inner contour of the cranial structure feature long.
该可选的实施例中，第一交点包括第一子交点和第二子交点，第二交点包 括第三子交点和第四子交点。其中，第一子交点与第三子交点的距离小于与第四子交点的距离。基于该第一交点与该第二交点，确定颅脑结构特征对应的双顶径几何参数，具体的：连接第一子交点与第四子交点所形成的第一线段，作为颅脑结构特征对应的双顶径几何参数，即双顶径长度；或者，连接第二子交点与第三子交点所形成的第二线段，作为颅脑结构特征对应的双顶径几何参数；或者，获取第一线段与第二线段的均值，作为颅脑结构特征对应的双顶径几何参数。这样通过提供多种颅脑结构特征的双顶径长度的获取方式，能够提高双顶径长度的获取可能性以及准确性。In this optional embodiment, the first intersection includes a first sub-intersection and a second sub-intersection, and the second intersection includes a third sub-intersection and a fourth sub-intersection. Wherein, the distance between the first sub-intersection and the third sub-intersection is smaller than the distance from the fourth sub-intersection. Based on the first intersection point and the second intersection point, determine the biparietal diameter geometric parameter corresponding to the cranial structure feature, specifically: connect the first line segment formed by the first sub-intersection and the fourth sub-intersection as the cranial structure feature The corresponding biparietal diameter geometric parameter, that is, the biparietal diameter length; or, the second line segment formed by connecting the second sub-intersection and the third sub-intersection, as the biparietal diameter geometric parameter corresponding to the structural features of the brain; or, obtain the first The mean value of the first line segment and the second line segment is used as the geometric parameter of the biparietal diameter corresponding to the structural features of the brain. In this way, the possibility and accuracy of obtaining the length of the biparietal diameter can be improved by providing ways of obtaining the length of the biparietal diameter of various craniocerebral structural features.
可见，该可选的实施例通过获取颅脑结构特征的内、外轮廓的周长，以及颅脑结构特征的脑中线对应的中垂线与内、外轮廓的交点，能够实现颅脑结构特征的头围周长以及双顶径长度的获取。It can be seen that in this optional embodiment, the perimeter of the inner and outer contours of the structural features of the skull, and the intersection of the midline corresponding to the midline of the brain and the inner and outer contours of the structural features of the skull, can realize the structural features of the skull. The perimeter of the head circumference and the length of the double parietal diameter were obtained.
在又一个可选的实施例中，在执行完毕步骤103之后，该方法还可以包括以下步骤：In yet another optional embodiment, after performing step 103, the method may further include the following steps:
根据胎儿超声图像的至少一个结构特征的几何参数确定该胎儿超声图像对应的胎儿的孕周。The gestational age of the fetus corresponding to the fetal ultrasound image is determined according to the geometric parameter of at least one structural feature of the fetal ultrasound image.
该可选的实施例中，可以根据胎儿超声图像的每个结构特征的几何参数单独确定该胎儿超声图像对应的胎儿的孕周，也可以根据多个结构特征的几何参数共同确定该胎儿超声图像对应的胎儿的孕周。In this optional embodiment, the gestational age of the fetus corresponding to the fetal ultrasound image may be determined independently according to the geometric parameters of each structural feature of the fetal ultrasound image, or the fetal ultrasound image may be jointly determined according to the geometric parameters of multiple structural features The corresponding gestational age of the fetus.
可见，该可选的实施例在获取到胎儿超声图像的结构特征的几何参数之后，能够根据胎儿超声图像的结构特征的几何参数单独或者共同确定该胎儿超声图像对应的胎儿的孕周，从而提高胎儿孕周的确定准确性。It can be seen that in this optional embodiment, after the geometric parameters of the structural features of the fetal ultrasound image are acquired, the gestational age of the fetus corresponding to the fetal ultrasound image can be determined individually or jointly according to the geometric parameters of the structural features of the fetal ultrasound image, thereby improving The accuracy of determination of fetal gestational age.
在又一个可选的实施例中，在执行步骤101之前，该方法还可以包括以下步骤：In yet another optional embodiment, before performing step 101, the method may further include the following steps:
基于获取到的样本结构特征集合包括的每个样本结构特征以及该样本结构特征的测量要素训练确定出的初始特征检测模型，得到训练后的特征检测模型，并确定训练后的特征检测模型为步骤101中的确定出的特征检测模型。Based on each sample structural feature included in the obtained sample structural feature set and the measurement elements of the sample structural feature, train the determined initial feature detection model, obtain the trained feature detection model, and determine the trained feature detection model as a step The determined feature detection model in 101.
该可选的实施例中，所有样本结构特征包括样本颅脑结构特征、样本腹围结构特征、样本小脑结构特征、样本第三脑室结构特征、样本侧脑室结构特征、样本透明隔结构特征、样本肱骨结构特征以及样本股骨结构特征中的至少一种，该可选的实施例不做限定。In this optional embodiment, all the structural features of the samples include the structural features of the skull of the sample, the structural features of the abdominal circumference of the sample, the structural features of the cerebellum of the sample, the structural features of the third ventricle of the sample, the structural features of the lateral ventricle of the sample, the structural At least one of the structural features of the humerus and the structural features of the sample femur is not limited to this optional embodiment.
该可选的实施例中，进一步的，每种样本结构特征均包括正样本结构特征和负样本结构特征，其中，正样本结构特征为正常的样本结构特征和负样本结构特征为异常的样本结构特征。这样通过将正常的样本结构特征以及异常的样本结构特征进行训练，有利于获取到功能更丰富的特征检测模型，从而有利于后续当胎儿超声图像存在异常结构特征时，能够将正常结构特征和异常结构特征检测出来，从而有利于获取到更全面的胎儿发育信息，进而提高胎儿的生长发育情况的确定准确性。In this optional embodiment, further, each sample structure feature includes a positive sample structure feature and a negative sample structure feature, wherein the positive sample structure feature is a normal sample structure feature and the negative sample structure feature is an abnormal sample structure feature. In this way, by training the normal sample structural features and abnormal sample structural features, it is beneficial to obtain a feature detection model with more functions, which is conducive to the subsequent use of abnormal structural features in fetal ultrasound images. Structural features are detected, which is conducive to obtaining more comprehensive fetal development information, thereby improving the accuracy of determining the growth and development of the fetus.
该可选的实施例中，初始特征检测模型可以包括目标检测模型、实例分割模型以及语义分割模型等中的一种或者多种模型的组合，该可选的实施例不做限定。此时，训练后的特征检测模型与初始特征检测模型对应，也即训练后的特征检测模型可以包括训练后的目标检测模型、训练后的实例分割模型以及训练后的语义分割模型等能够获取到胎儿超声图像的特征信息中的一种或者多种 模型的组合，该可选的实施例不做限定。In this optional embodiment, the initial feature detection model may include a target detection model, an instance segmentation model, a semantic segmentation model, and a combination of one or more models, which are not limited in this optional embodiment. At this point, the trained feature detection model corresponds to the initial feature detection model, that is, the trained feature detection model may include the trained target detection model, the trained instance segmentation model, and the trained semantic segmentation model, etc. The optional embodiment is not limited to a combination of one or more models in the feature information of the fetal ultrasound image.
该可选的实施例中，不同的样本结构特征对应不同的测量要素，具体的：样本小脑结构特征对应的测量要素为样本小脑结构特征(小脑半球)最宽位置的连线，且该连线垂直于脑中线；样本第三脑室结构特征对应的测量要素为第三脑室空腔区域最宽区域的线段，且该线段垂直于脑中线；样本侧脑室结构特征对应的测量要素为侧脑室空腔最宽区域的线段，且该线段垂直于外接矩形最长边；样本透明隔结构特征对应的测量要素为透明隔空腔最宽区域的线段，且该线段垂直于外接矩形最长边；样本肱骨结构特征对应的测量要素为肱骨最宽区域的线段，且该线段垂直于外接矩形最长边；样本股骨结构特征对应的测量要素为股骨最宽区域的线段，且该线段垂直于外接矩形最长边。In this optional embodiment, different sample structural features correspond to different measurement elements, specifically: the measurement elements corresponding to the sample cerebellar structural features are the lines connecting the widest positions of the sample cerebellum structural features (cerebellar hemispheres), and the connecting line Perpendicular to the midline of the brain; the measurement element corresponding to the structural features of the third ventricle of the sample is the line segment of the widest region of the third ventricle cavity, and the line segment is perpendicular to the midline of the brain; the measurement element corresponding to the structural features of the lateral ventricle of the sample is the lateral ventricle cavity The line segment of the widest region, and the line segment is perpendicular to the longest side of the circumscribed rectangle; the measurement element corresponding to the transparent septum structure feature of the sample is the line segment of the widest region of the transparent septum cavity, and the line segment is perpendicular to the longest side of the circumscribed rectangle; the sample humerus The measurement element corresponding to the structural feature is the line segment of the widest region of the humerus, and the line segment is perpendicular to the longest side of the circumscribed rectangle; the measurement element corresponding to the structural feature of the sample femur is the line segment of the widest region of the femur, and the line segment is perpendicular to the longest side of the circumscribed rectangle. side.
可见，该可选的实施例通过预先训练符合本方案的特征检测模型，能够便于后续直接使用该特征检测模型，从而提高胎儿超声图像的结构特征的信息，进而提高胎儿超声图像的结构特征的几何参数的获取效率以及准确性；以及在训练特征检测模型的过程中，将对应结构特征的测量要素一起进行训练，能够提高特征检测模型的训练准确性以及可靠性，从而有利于进一步提高胎儿超声图像的结构特征的几何参数的获取效率以及准确性。It can be seen that this optional embodiment can facilitate the subsequent use of the feature detection model by pre-training a feature detection model that conforms to this scheme, thereby improving the information of the structural features of the fetal ultrasound image, and then improving the geometrical geometry of the structural features of the fetal ultrasound image. The efficiency and accuracy of parameter acquisition; and in the process of training the feature detection model, the measurement elements corresponding to structural features are trained together, which can improve the training accuracy and reliability of the feature detection model, which is conducive to further improving fetal ultrasound images. The acquisition efficiency and accuracy of the geometric parameters of the structural features.
实施例二Embodiment 2
请参阅图2，图2是本发明实施例公开的另一种胎儿结构特征的自动测量方法的流程示意图。其中，图2所描述的胎儿结构特征的自动测量方法可以应用于参数测量服务器中，其中，该参数测量服务器可以包括本地参数测量服务器或云参数测量服务器，本发明实施例不做限定。如图2所示，该胎儿结构特征的自动测量方法可以包括以下操作：Please refer to FIG. 2 , which is a schematic flowchart of another method for automatically measuring fetal structural features disclosed in an embodiment of the present invention. The automatic measurement method for fetal structural features described in FIG. 2 may be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention. As shown in FIG. 2, the automatic measurement method of fetal structural features may include the following operations:
201、将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析，并获取特征检测模型输出的分析结果，作为该胎儿超声图像的特征信息，该胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓。201. Input the acquired fetal ultrasound image into the determined feature detection model for analysis, and obtain an analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the fetal ultrasound image. The category of at least one structural feature of the image and the contour of at least one structural feature.
202、基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓。202. Perform a preprocessing operation on the contour of each structural feature based on the determined preprocessing manner, to obtain a preprocessed contour of the structural feature.
本发明的实施例中，预处理操作包括平滑处理操作。进一步的，该预处理操作还可以包括图像增强、图像滤波、图像分割操作中的至少一种，该可选的实施例不做限定。In the embodiment of the present invention, the preprocessing operation includes a smoothing operation. Further, the preprocessing operation may further include at least one of image enhancement, image filtering, and image segmentation operations, which are not limited in this optional embodiment.
需要说明的是，步骤202和步骤203的发生顺序没有先后之分，即步骤202可以发生在步骤203之后，也可以发生在步骤203之前，还可以和步骤203同时发生，本发明实施例不做限定。It should be noted that there is no sequence of occurrence of step 202 and step 203, that is, step 202 may occur after step 203, or before step 203, or may occur at the same time as step 203, which is not implemented in this embodiment of the present invention. limited.
可见，本发明实施例在获取到胎儿超声图像的特征信息之后，进一步对胎儿超声图像的结构特征的轮廓执行预处理操作，例如：平滑处理操作，能够将结构特征的轮廓中的噪声滤除，获取到清晰的结构特征的轮廓，从而有利于提高结构特征的几何参数的获取效率以及精准性。It can be seen that, after acquiring the feature information of the fetal ultrasound image, the embodiment of the present invention further performs a preprocessing operation on the contour of the structural feature of the fetal ultrasound image, such as a smoothing operation, which can filter out the noise in the contour of the structural feature, A clear outline of the structural feature is obtained, thereby helping to improve the efficiency and accuracy of obtaining the geometrical parameters of the structural feature.
203、根据每个结构特征的类别确定与该结构特征的轮廓相匹配的测量方式。203. Determine, according to the category of each structural feature, a measurement mode that matches the contour of the structural feature.
204、根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数，此时，步骤204中的每个结构特征的轮廓均为预处理后的结构特征的轮廓。204. Perform a measurement operation on the contour of the structural feature according to the measurement method of each structural feature to obtain the geometric parameters of the structural feature. At this time, the contour of each structural feature in step 204 is the preprocessed structural feature. contour.
本发明实施例中，针对步骤201、步骤203以及步骤204的其他描述请参阅实施例一中的详细描述，本发明实施例不再赘述。In this embodiment of the present invention, for other descriptions of step 201, step 203, and step 204, please refer to the detailed description in Embodiment 1, and details are not repeated in this embodiment of the present invention.
在一个可选的实施例中，在执行完毕步骤201之后，该方法还可以包括以下步骤：In an optional embodiment, after performing step 201, the method may further include the following steps:
根据胎儿超声图像的特征信息判断每个结构特征的轮廓是否满足确定出的预处理条件；Determine whether the contour of each structural feature meets the determined preprocessing conditions according to the feature information of the fetal ultrasound image;
当判断出每个结构特征均不满足预处理条件结果时，触发执行步骤204；When it is determined that each structural feature does not meet the preprocessing condition result, triggering to execute step 204;
当判断出所有结构特征的轮廓中存在满足预处理条件的轮廓时，从所有结构特征中筛选满足预处理条件的所有目标结构特征，并触发执行步骤202，此时，步骤202骤中的所有结构特征包括所有目标结构特征。需要说明的是，当存在目标结构特征，即存在需要执行预处理操作的结构特征时，可以直接对其他不需要执行预处理操作的结构特征的轮廓执行确定与之匹配的测量方式的操作，也可以待对目标结构特征执行完毕预处理操作之后，和预处理后的目标结构特征一并执行确定与之匹配的测量方式的操作，该可选的实施例不做限定。When it is determined that there are contours that meet the preprocessing conditions in the contours of all the structural features, all target structural features that meet the preprocessing conditions are screened from all the structural features, and the execution of step 202 is triggered. At this time, all the structures in step 202 Features include all target structural features. It should be noted that when there is a target structural feature, that is, there is a structural feature that needs to be preprocessed, the operation of determining the matching measurement method can be directly performed on the contours of other structural features that do not need to be preprocessed. After the preprocessing operation is performed on the target structure feature, the operation of determining a matching measurement mode may be performed together with the preprocessed target structure feature, which is not limited in this optional embodiment.
可见，该可选的实施例在获取到胎儿超声图像的特征信息之后，进一步判断胎儿超声图像的结构特征是否满足预处理条件，若不满足预处理条件，则直接执行结构特征的轮廓的测量操作，能够提高结构特征的几何参数的测量效率，从而提高胎儿的生长发育情况的确定效率；若满足预处理条件，对满足预处理条件的结构特征的轮廓执行预处理操作，能够获取到准确可靠的结构特征的轮廓，从而提高结构特征的几何参数的测量准确性以及可靠性，从而进一步提高胎儿的生长发育情况的确定准确性以及可靠性。It can be seen that in this optional embodiment, after acquiring the feature information of the fetal ultrasound image, it is further judged whether the structural feature of the fetal ultrasound image satisfies the preprocessing condition, and if the preprocessing condition is not met, the measurement operation of the contour of the structural feature is directly performed. , which can improve the measurement efficiency of the geometric parameters of structural features, thereby improving the efficiency of determining the growth and development of the fetus; if the preprocessing conditions are met, the preprocessing operation is performed on the contours of the structural features that meet the preprocessing conditions, and accurate and reliable information can be obtained. The contour of the structural feature improves the measurement accuracy and reliability of the geometric parameters of the structural feature, thereby further improving the accuracy and reliability of the determination of the growth and development of the fetus.
在另一个可选的实施例中，根据胎儿超声图像的特征信息判断每个结构特征的轮廓是否满足确定出的预处理条件，可以包括：In another optional embodiment, judging whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image may include:
根据每个结构特征的类别判断该结构特征是否为确定出的需要执行预处理操作的结构特征的类别，当判断结果为是时，确定该结构特征满足确定出的预处理条件；或者，According to the category of each structural feature, it is determined whether the structural feature is the determined category of the structural feature that needs to perform a preprocessing operation, and when the judgment result is yes, it is determined that the structural feature satisfies the determined preprocessing condition; or,
根据每个结构特征的轮廓判断该结构特征的轮廓是否满足确定出的参数测量条件，当判断结果为否时，确定该结构特征满足确定出的预处理条件。According to the contour of each structural feature, it is determined whether the contour of the structural feature satisfies the determined parameter measurement condition, and when the judgment result is no, it is determined that the structural feature satisfies the determined preprocessing condition.
该可选的实施例中，进一步的，当判断出每个结构特征的类别为需要执行预处理操作的结构特征的类别以及判断出该结构特征的轮廓不满足确定出的参数测量条件时，才确定该结构特征满足确定出的预处理条件。这样不仅能够进一步提高胎儿超声图像的结构特征的轮廓需要执行预处理操作的确定准确性，还能够丰富胎儿超声图像的结构特征的轮廓需要执行预处理操作的确定方式。In this optional embodiment, further, only when it is determined that the category of each structural feature is the category of the structural feature that needs to perform a preprocessing operation and it is determined that the contour of the structural feature does not meet the determined parameter measurement conditions. It is determined that the structural feature satisfies the determined preprocessing conditions. This can not only further improve the accuracy of determining that the contour of the structural feature of the fetal ultrasound image needs to perform the preprocessing operation, but also enrich the way of determining that the contour of the structural feature of the fetal ultrasound image needs to perform the preprocessing operation.
该可选的实施例中，需要执行预处理操作的结构特征的类别可以包括颅脑结构特征、腹围结构特征、小脑结构特征以及心脏结构特征中的一种或者多种，该可选的实施例不做限定。不需要执行预处理操作的结构特征的类别可以包括侧脑室结构特征、股骨结构特征、肱骨结构特征、透明隔结构特征中的一种或者多种，该可选的实施例不做限定。In this optional embodiment, the categories of structural features that need to perform preprocessing operations may include one or more of the structural features of the cranial brain, the structural features of the abdominal circumference, the structural features of the cerebellum, and the structural features of the heart. Examples are not limited. The categories of structural features that do not need to perform preprocessing operations may include one or more of lateral ventricle structural features, femur structural features, humerus structural features, and septum pellucidum structural features, which are not limited in this optional embodiment.
该可选的实施例中，当胎儿超声图像的结构特征的轮廓对应的轮廓信息(例如：分辨率、清晰度、对比度、噪声等)不满足参数测量条件时，确定该结构特征满足确定出的预处理条件。举例来说，当胎儿超声图像的结构特征的轮廓存在噪声和/或对比度低于确定出的对比度阈值(例如：90％)时，确定该结构 特征不满足参数测量条件，即需要对该结构特征的轮廓执行预处理操作。In this optional embodiment, when the contour information (for example, resolution, sharpness, contrast, noise, etc.) corresponding to the contour of the structural feature of the fetal ultrasound image does not satisfy the parameter measurement conditions, it is determined that the structural feature satisfies the determined condition. preprocessing conditions. For example, when the contour of the structural feature of the fetal ultrasound image has noise and/or the contrast is lower than a determined contrast threshold (for example: 90%), it is determined that the structural feature does not meet the parameter measurement conditions, that is, the structural feature needs to be The contour of the preprocessing operation is performed.
可见，该可选的实施例通过在判断出胎儿超声图像的结构特征的类别为需要执行预处理条件的结构特征的类别和/或该结构特征的轮廓不满足参数测量条件时，确定结构特征满足确定出的预处理条件，能够提高胎儿超声图像的结构特征需要执行预处理操作的确定准确性，还能够丰富胎儿超声图像的结构特征的轮廓需要执行预处理操作的确定方式，有利于获取到满足要求的结构特征的轮廓，从而有利于进一步提高结构特征的几何参数的测量准确性以及可靠性，进而进一步提高胎儿的生长发育情况确定准确性以及可靠性。It can be seen that in this optional embodiment, when it is determined that the category of the structural feature of the fetal ultrasound image is the category of the structural feature that needs to execute the preprocessing condition and/or the contour of the structural feature does not satisfy the parameter measurement condition, it is determined that the structural feature satisfies the The determined preprocessing conditions can improve the accuracy of determining that the structural features of the fetal ultrasound image need to be preprocessed, and can also enrich the way that the contours of the structural features of the fetal ultrasound image need to be preprocessed. The profile of the required structural features is beneficial to further improve the measurement accuracy and reliability of the geometric parameters of the structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
在又一个可选的实施例中，每个结构特征的轮廓均对应多个节点。以及，基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，可以包括：In yet another optional embodiment, the contour of each structural feature corresponds to multiple nodes. And, based on the determined preprocessing method, a preprocessing operation is performed on the contour of each structural feature to obtain the contour of the preprocessed structural feature, which may include:
获取每个结构特征的轮廓对应的圆弧半径；Obtain the arc radius corresponding to the contour of each structural feature;
当每个结构特征的轮廓对应的圆弧半径大于等于确定出的圆弧半径阈值(例如：5mm)时，从该结构特征对应的所有节点中选取预设数量的目标节点，并按照每相邻两个节点进行连接的方式将每个结构特征对应的所有目标节点依次连接起来，得到预处理后的该结构特征的轮廓；When the arc radius corresponding to the contour of each structural feature is greater than or equal to the determined arc radius threshold (for example: 5mm), select a preset number of target nodes from all the nodes corresponding to the structural feature, and select a preset number of target nodes according to each adjacent The way of connecting the two nodes connects all the target nodes corresponding to each structural feature in turn to obtain the preprocessed outline of the structural feature;
当每个结构特征的轮廓对应的圆弧半径不大于等于确定出的圆弧半径阈值时，按照每相邻两个节点进行连接的方式将每个结构特征对应的所有节点依次连接起来，得到预处理后的该结构特征的轮廓。When the arc radius corresponding to the contour of each structural feature is not greater than or equal to the determined arc radius threshold, connect all the nodes corresponding to each structural feature in turn according to the method of connecting every two adjacent nodes to obtain a predetermined The processed outline of the structural feature.
该可选的实施例中，当结构特征的轮廓存在多个圆弧和/或轮廓的曲率大于等于确定出的曲率阈值时，对该结构特征的轮廓进行分段执行预处理操作。具体的：当结构特征的轮廓存在多个圆弧时，将分别对该结构特征的多个圆弧中每个圆弧执行预处理操作；当结构特征的轮廓的曲率大于等于曲率阈值时，将结构特征的轮廓等间隔或非等间隔分成多段，并分别对每个每段轮廓执行预处理操作。这样在结构特征的轮廓存在多个圆弧和/或轮廓的曲率较大时，通过对结构特征的轮廓分段执行预处理操作，能够提高结构特征的轮廓的预处理效率以及准确性，从而有利于进一步提高胎儿超声图像的结构特征的几何参数的测量准确性以及可靠性。可见，该可选的实施例通过根据胎儿超声图像的结构特征的圆弧半径的大小选择不同的预处理方式，不仅能够实现结构特征的预处理，还能够提高结构特征的预处理效率以及准确性，从而提高结构特征的几何参数的测量准确性，进而进一步提高胎儿超声图像的对应的胎儿的生长发育情况的确定准确性。In this optional embodiment, when the contour of the structural feature has multiple arcs and/or the curvature of the contour is greater than or equal to the determined curvature threshold, the contour of the structural feature is segmented to perform a preprocessing operation. Specifically: when there are multiple arcs in the contour of the structural feature, the preprocessing operation will be performed on each of the multiple arcs of the structural feature; when the curvature of the contour of the structural feature is greater than or equal to the curvature threshold, the The contours of structural features are divided into multiple segments at equal or unequal intervals, and preprocessing operations are performed on each segment of the contour separately. In this way, when the contour of the structural feature has multiple arcs and/or the curvature of the contour is large, by performing preprocessing operations on the contour segment of the structural feature, the preprocessing efficiency and accuracy of the contour of the structural feature can be improved, so that there is a It is beneficial to further improve the measurement accuracy and reliability of the geometric parameters of the structural features of the fetal ultrasound image. It can be seen that this optional embodiment can not only realize the preprocessing of structural features, but also improve the preprocessing efficiency and accuracy of structural features by selecting different preprocessing methods according to the size of the arc radius of the structural features of the fetal ultrasound image. , thereby improving the measurement accuracy of the geometric parameters of the structural features, and further improving the determination accuracy of the fetal growth and development situation corresponding to the fetal ultrasound image.
在又一个可选的实施例中，基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，可以包括：In yet another optional embodiment, performing a preprocessing operation on the contour of each structural feature based on the determined preprocessing method, to obtain the contour of the preprocessed structural feature, which may include:
基于确定出的B样条曲线拟合方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓。Based on the determined B-spline curve fitting method, a preprocessing operation is performed on the contour of each structural feature to obtain the contour of the preprocessed structural feature.
该可选的实施例中，B样条曲线拟合方式可以包括插值拟合方式和/或近似拟合方式，该可选的实施例不做限定。In this optional embodiment, the B-spline curve fitting method may include an interpolation fitting method and/or an approximate fitting method, which is not limited in this optional embodiment.
可见，该可选的实施例通过提供B样条曲线拟合方式，不仅能够实现胎儿超声图像的结构特征的轮廓的平滑拟合，还能够丰富结构特征的轮廓的平滑拟合方式。It can be seen that, by providing a B-spline curve fitting method, this optional embodiment can not only achieve smooth fitting of the contour of the structural feature of the fetal ultrasound image, but also enrich the smooth fitting method of the contour of the structural feature.
可见，实施图2所描述的胎儿结构特征的自动测量方法能够通过自动获取 胎儿的结构特征的类别和轮廓，并根据对应的测量方式测量结构特征的轮廓，无需通过手动获取胎儿结构特征的几何参数，能够快速获取到高精度的胎儿结构特征的测量结果，从而准确确定胎儿的生长发育情况；以及通过将胎儿超声图像输入特征检测模型，能够提高胎儿的结构特征的测量结果的获取效率。此外，还能够将结构特征的轮廓中的噪声滤除，获取到清晰的结构特征的轮廓，从而有利于提高结构特征的几何参数的获取效率以及精准性。It can be seen that implementing the automatic measurement method of fetal structural features described in FIG. 2 can automatically obtain the categories and contours of fetal structural features, and measure the contours of structural features according to the corresponding measurement methods, without manually obtaining the geometric parameters of fetal structural features. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results. In addition, the noise in the contour of the structural feature can be filtered out, and a clear contour of the structural feature can be obtained, thereby helping to improve the efficiency and accuracy of obtaining geometric parameters of the structural feature.
实施例三Embodiment 3
请参阅图3，图3是本发明实施例公开的一种胎儿结构特征的自动测量装置的结构示意图。其中，图3所描述的胎儿结构特征的自动测量装置可以应用于参数测量服务器中，其中，该参数测量服务器可以包括本地参数测量服务器或云参数测量服务器，本发明实施例不做限定。如图3所示，该胎儿结构特征的自动测量装置可以包括分析模块301、获取模块302以及测量模块303，其中：Please refer to FIG. 3 , which is a schematic structural diagram of an automatic measuring device for fetal structural features disclosed in an embodiment of the present invention. 3 can be applied to a parameter measurement server, wherein the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention. As shown in FIG. 3 , the automatic measuring device for fetal structural features may include an analysis module 301, an acquisition module 302 and a measurement module 303, wherein:
分析模块301，将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析。The analysis module 301 inputs the acquired fetal ultrasound image into the determined feature detection model for analysis.
获取模块302，用于获取特征检测模型输出的分析结果，作为该胎儿超声图像的特征信息，该胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓。The acquisition module 302 is configured to acquire the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the category of at least one structural feature of the fetal ultrasound image and the at least one structural feature of the fetal ultrasound image. contour.
本发明实施例中，可选的，所有结构特征包括颅脑结构特征、腹围结构特征以及其他结构特征中的至少一种，其他结构特征包括小脑结构特征、第三脑室结构特征、侧脑室结构特征、透明隔结构特征、肱骨结构特征以及股骨结构特征中的至少一种。In this embodiment of the present invention, optionally, all structural features include at least one of cranial brain structural features, abdominal circumference structural features, and other structural features, and other structural features include cerebellar structural features, third ventricle structural features, and lateral ventricle structural features at least one of a feature, a septum pellucida feature, a humerus feature, and a femur feature.
测量模块303，用于根据每个结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。The measurement module 303 is configured to determine a measurement mode that matches the contour of the structural feature according to the category of each structural feature, and perform a measurement operation on the contour of the structural feature according to the measurement mode of each structural feature, to obtain the contour of the structural feature. Geometric parameters.
可见，实施图3所描述的胎儿结构特征的自动测量装置能够通过自动获取胎儿的结构特征的类别和轮廓，并根据对应的测量方式测量结构特征的轮廓，无需通过手动获取胎儿结构特征的几何参数，能够快速获取到高精度的胎儿结构特征的测量结果，从而准确确定胎儿的生长发育情况；以及通过将胎儿超声图像输入特征检测模型，能够提高胎儿的结构特征的测量结果的获取效率。It can be seen that the automatic measuring device implementing the fetal structural feature described in FIG. 3 can automatically obtain the category and contour of the fetal structural feature, and measure the contour of the structural feature according to the corresponding measurement method, without manually obtaining the geometrical parameters of the fetal structural feature. , can quickly obtain high-precision measurement results of fetal structural features, so as to accurately determine the growth and development of the fetus; and by inputting fetal ultrasound images into the feature detection model, it can improve the acquisition efficiency of fetal structural features measurement results.
在一个可选的实施例中，如图3所示，测量模块303根据每个结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的方式具体为：In an optional embodiment, as shown in FIG. 3 , the measurement module 303 determines a measurement method that matches the outline of each structural feature according to the category of each structural feature, and determines the measurement method for the structure according to the measurement method of each structural feature. The outline of the feature is measured, and the method to obtain the geometric parameters of the structural feature is as follows:
当上述结构特征为颅脑结构特征时，基于颅脑结构特征的内轮廓与该颅脑结构特征的外轮廓，确定该颅脑结构特征对应的几何参数，该颅脑结构特征的轮廓包括外轮廓和内轮廓；When the above-mentioned structural feature is a cranial structure feature, the geometric parameters corresponding to the cranial structure feature are determined based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature, and the contour of the cranial structure feature includes the outer contour and inner contour;
当上述结构特征为腹围结构特征时，获取腹围结构特征的轮廓对应的周长，作为腹围结构特征的几何参数；When the above-mentioned structural feature is the abdominal circumference structural feature, the perimeter corresponding to the contour of the abdominal circumference structural feature is obtained as the geometric parameter of the abdominal circumference structural feature;
当上述结构特征为其他结构特征时，基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数，图形拟合方式包括最小外接矩形拟合方式和/或椭圆拟合方式。When the above-mentioned structural features are other structural features, fit the contour of each other structural feature based on the determined graph fitting method to obtain the target node of the other structural feature, and determine the target node of the other structural feature based on the target node of each other structural feature. Geometric parameters of other structural features, and graph fitting methods include minimum circumscribed rectangle fitting methods and/or ellipse fitting methods.
该可选的实施例中，可选的，测量模块303基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数的方式具体为：In this optional embodiment, optionally, the measurement module 303 fits the contour of each other structural feature based on the determined graphic fitting method, obtains the target node of the other structural feature, and determines the target node of each other structural feature based on the determined graph fitting method For the target node, the method of determining the geometric parameters of the other structural features is as follows:
当图形拟合方式为最小外接矩形拟合方式时，基于确定出的外接矩形拟合与其他结构特征的轮廓，当外接矩形的其中一组对边均与其他结构特征的轮廓存在目标交点时，停止拟合其他结构特征的轮廓，得到其他结构特征的轮廓的最小外接矩形，并根据该其他结构特征的类别确定该其他结构特征的扫查方向，并基于该扫查方向对该其他结构特征的轮廓执行扫查操作，得到该其他结构特征的轮廓的多条扫查线段，并从所有扫查线段中选择距离值最大的扫查线段，作为该其他结构特征的几何参数。When the graph fitting mode is the minimum circumscribed rectangle fitting mode, based on the determined circumscribed rectangle fitting and the contours of other structural features, when a set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, Stop fitting the contours of other structural features, obtain the minimum circumscribed rectangle of the contours of other structural features, determine the scanning direction of the other structural features according to the category of the other structural features, and determine the scanning direction of the other structural features based on the scanning direction. The contour scan operation is performed to obtain multiple scan line segments of the contour of the other structural feature, and the scan line segment with the largest distance value is selected from all the scan line segments as the geometric parameters of the other structural feature.
该可选的实施方式中，扫查方向包括第一扫查方向或第二扫查方向中的其中一种，其中，该第一扫查方向用于表示从平行于最小外接矩形的短边对该其他结构特征的轮廓进行扫查，该第二扫查方向用于表示从平行于最小外接矩形的长边对该其他结构特征的轮廓进行扫查，例如：侧脑室增宽结构特征的扫查方向为平行于最小外接矩形的短边，左心房结构特征的扫查方向为平行于最小外接矩形的长边。又可选的，该目标交点为在外接矩形无线逼近其他结构特征的轮廓的过程中，当外接矩形的其中一组对边一旦均与其他结构特征的轮廓存在目标交点时，就停止拟合其他结构特征的轮廓，这样能够提高结构特征的扫查方向的确定准确性，从而进一步提高结构特征的几何参数的获取准确性、可靠性以及效率。In this optional implementation manner, the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to represent a pair of short sides parallel to the smallest circumscribed rectangle. The contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle, for example: scanning of the structural feature of lateral ventricle enlargement The direction is parallel to the short side of the smallest circumscribed rectangle, and the scanning direction of the structural features of the left atrium is parallel to the long side of the smallest circumscribed rectangle. Alternatively, the target intersection point is in the process of wirelessly approximating the contours of other structural features by the circumscribed rectangle, when one set of opposite sides of the circumscribed rectangle has a target intersection point with the contours of other structural features, stop fitting other structural features. The outline of the structural feature can improve the accuracy of determining the scanning direction of the structural feature, thereby further improving the accuracy, reliability and efficiency of obtaining geometric parameters of the structural feature.
该可选的实施例中，可选的，测量模块303基于确定出的图形拟合方式拟合每个其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个其他结构特征的目标节点，确定该其他结构特征的几何参数的方式具体为：In this optional embodiment, optionally, the measurement module 303 fits the contour of each other structural feature based on the determined graphic fitting method, obtains the target node of the other structural feature, and determines the target node of each other structural feature based on the determined graph fitting method For the target node, the method of determining the geometric parameters of the other structural features is as follows:
当图形拟合方式为椭圆拟合方式时，基于确定出的椭圆拟合算法拟合其他结构特征的轮廓，得到拟合后的其他结构特征的轮廓，拟合后的其他结构特征的轮廓的形状为椭圆形；When the graph fitting method is the ellipse fitting method, the contours of other structural features are fitted based on the determined ellipse fitting algorithm to obtain the contours of the fitted other structural features and the shape of the fitted contours of the other structural features. is oval;
根据其他结构特征的类别确定其他结构特征的目标扫查方向；Determine the target scanning direction of other structural features according to the categories of other structural features;
基于目标扫查方向对其他结构特征的轮廓执行扫查操作，得到其他结构特征的轮廓的多条目标扫查线段，并从所有目标扫查线段中选择距离值最大的目标扫查线段，作为其他结构特征的几何参数；Perform the scanning operation on the contours of other structural features based on the scanning direction of the target, obtain multiple target scanning line segments of the contours of other structural features, and select the target scanning line segment with the largest distance value from all the target scanning line segments as the other scanning line segments. geometric parameters of structural features;
其中，目标扫查方向包括第一目标扫查方向或第二目标扫查方向中的其中一种，其中，第一目标扫查方向用于表示从平行于拟合后的其他结构特征的轮廓的短轴对其他结构特征的轮廓进行扫查，第二目标扫查方向用于表示从平行于拟合后的其他结构特征的轮廓的长轴对其他结构特征的轮廓进行扫查。Wherein, the target scanning direction includes one of the first target scanning direction or the second target scanning direction, wherein the first target scanning direction is used to represent the direction from the contour parallel to the other structural features after fitting. The short axis scans the contours of other structural features, and the second target scanning direction is used to indicate that the contours of other structural features are scanned from the long axis parallel to the contours of the other structural features after fitting.
可见，实施图3所描述的装置还能够根据不同结构特征，采取不同几何参数的测量方式测量对应结构特征的几何参数，不仅能够实现胎儿超声图像的结构特征的几何参数的获取，还能够提高胎儿超声图像的结构特征的几何参数的获取准确性以及效率；通过最小外接矩形拟合结构特征，能够实现结构特征的几何参数的获取，以及不同类别的结构特征选择不同的扫查方向，能够提高结构特征的几何参数的获取准确性、可靠性以及效率；以及通过椭圆拟合方式和/或最小外接矩形拟合方式等方式获取结构特征的几何参数，能够丰富结构特征的几何参数获取方式，以及提高结构特征的几何参数获取可能性以及准确性。It can be seen that implementing the device described in FIG. 3 can also measure the geometric parameters of the corresponding structural features by measuring different geometric parameters according to different structural features. Accuracy and efficiency of obtaining geometric parameters of structural features of ultrasound images; by fitting the structural features with the smallest circumscribed rectangle, the geometric parameters of structural features can be obtained, and different types of structural features can be selected in different scanning directions, which can improve the structural features. Accuracy, reliability and efficiency of obtaining geometric parameters of features; and obtaining geometric parameters of structural features by means of ellipse fitting and/or least circumscribed rectangle fitting, which can enrich the ways of obtaining geometric parameters of structural features, and improve the Possibility and accuracy of geometric parameter acquisition of structural features.
在另一个可选的实施例中，如图3所示，测量模块303基于颅脑结构特征的内轮廓与该颅脑结构特征的外轮廓，确定该颅脑结构特征对应的几何参数的方式具体为：In another optional embodiment, as shown in FIG. 3 , the measuring module 303 determines a specific way of determining the geometric parameter corresponding to the cranial structure feature based on the inner contour of the cranial structure feature and the outer contour of the cranial structure feature for:
获取颅脑结构特征的内轮廓的第一周长与该颅脑结构特征的外轮廓的第二周长，并基于该第一周长和第该二周长，确定颅脑结构特征对应的头围几何参数；Obtain the first perimeter of the inner contour of the cranial structural feature and the second perimeter of the outer contour of the cranial structural feature, and determine the head corresponding to the cranial structural feature based on the first perimeter and the second perimeter surrounding geometric parameters;
确定颅脑结构特征的脑中线对应的中垂线与颅脑结构特征的外轮廓的第一交点以及该中垂线与颅脑结构特征的内轮廓的第二交点，并基于该第一交点与该第二交点，确定颅脑结构特征对应的双顶径几何参数。Determine the first intersection point of the midline corresponding to the midline of the cranial structure feature and the outer contour of the cranial structure feature and the second intersection point of the perpendicular line and the inner contour of the craniocerebral structure feature, and based on the first intersection point and The second intersection point determines the biparietal diameter geometric parameters corresponding to the structural features of the skull.
可见，实施图3所描述的装置还能够通过获取颅脑结构特征的内、外轮廓的周长，以及颅脑结构特征的脑中线对应的中垂线与内、外轮廓的交点，能够实现颅脑结构特征的头围周长以及双顶径长度的获取。It can be seen that implementing the device described in FIG. 3 can also obtain the perimeter of the inner and outer contours of the structural features of the skull and the intersection of the mid-perpendicular line corresponding to the midline of the brain and the inner and outer contours of the structural features of the skull. Acquisition of head circumference and biparietal diameter length for structural features of the brain.
在又一个可选的实施例中，如图4所示，该装置还可以包括预处理模块304，其中：In yet another optional embodiment, as shown in FIG. 4 , the apparatus may further include a preprocessing module 304, wherein:
预处理模块304，用于在获取模块302获取特征检测模型输出的分析结果，作为胎儿超声图像的特征信息之后，基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，并触发测量模块303执行的根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作，每个结构特征的轮廓均为预处理后的结构特征的轮廓，该预处理操作包括预处理操作。The preprocessing module 304 is configured to, after the acquisition module 302 acquires the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, performs a preprocessing operation on the contour of each structural feature based on the determined preprocessing method to obtain a preprocessing operation. The contour of the processed structural feature, and trigger the measurement module 303 to perform a measurement operation on the contour of the structural feature according to the measurement method of each structural feature, and obtain the operation of the geometric parameters of the structural feature, and the contour of each structural feature is is the contour of the preprocessed structural features, and the preprocessing operation includes a preprocessing operation.
可见，实施图4所描述的装置能够在获取到胎儿超声图像的特征信息之后，进一步对胎儿超声图像的结构特征的轮廓执行预处理操作，例如：平滑处理操作，能够将结构特征的轮廓中的噪声滤除，获取到清晰的结构特征的轮廓，从而有利于提高结构特征的几何参数的获取效率以及精准性。It can be seen that, implementing the device described in FIG. 4 can further perform a preprocessing operation on the contour of the structural feature of the fetal ultrasound image after acquiring the feature information of the fetal ultrasound image, such as a smoothing operation, which can convert the contour of the structural feature into a Noise filtering is used to obtain clear outlines of structural features, which is beneficial to improve the efficiency and accuracy of geometric parameter acquisition of structural features.
在又一个可选的实施例中，可选的，每个结构特征的轮廓均对应多个节点；以及，如图4所示，预处理模块304基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的方式具体为：In yet another optional embodiment, optionally, the outline of each structural feature corresponds to multiple nodes; and, as shown in FIG. 4 , the preprocessing module 304 analyzes each structural feature based on the determined preprocessing method. The preprocessing operation is performed on the contour of , and the method of obtaining the contour of the preprocessed structural feature is as follows:
获取每个结构特征的轮廓对应的圆弧半径；Obtain the arc radius corresponding to the contour of each structural feature;
当每个结构特征的轮廓对应的圆弧半径大于等于确定出的圆弧半径阈值时，从该结构特征对应的所有节点中选取预设数量的目标节点，并按照每相邻两个节点进行连接的方式将每个结构特征对应的所有目标节点依次连接起来，得到预处理后的该结构特征的轮廓；When the arc radius corresponding to the contour of each structural feature is greater than or equal to the determined arc radius threshold, select a preset number of target nodes from all nodes corresponding to the structural feature, and connect each adjacent two nodes. Connect all the target nodes corresponding to each structural feature in turn to obtain the preprocessed contour of the structural feature;
当每个结构特征的轮廓对应的圆弧半径不大于等于确定出的圆弧半径阈值时，按照每相邻两个节点进行连接的方式将每个结构特征对应的所有节点依次连接起来，得到预处理后的该结构特征的轮廓。When the arc radius corresponding to the contour of each structural feature is not greater than or equal to the determined arc radius threshold, connect all the nodes corresponding to each structural feature in turn according to the method of connecting every two adjacent nodes to obtain a predetermined The processed outline of the structural feature.
可见，实施图4所描述的装置能够通过根据胎儿超声图像的结构特征的圆弧半径的大小选择不同的预处理方式，不仅能够实现结构特征的预处理，还能够提高结构特征的预处理效率以及准确性，从而提高结构特征的几何参数的测量准确性，进而进一步提高胎儿超声图像的对应的胎儿的生长发育情况的确定准确性。It can be seen that, by implementing the device described in FIG. 4, different preprocessing methods can be selected according to the size of the arc radius of the structural features of the fetal ultrasound image, which can not only realize the preprocessing of the structural features, but also can improve the preprocessing efficiency of the structural features and Therefore, the measurement accuracy of the geometric parameters of the structural features is improved, and the determination accuracy of the fetal growth and development situation corresponding to the fetal ultrasound image is further improved.
在又一个可选的实施例中，如图4所示，该装置还可以包括判断模块305以及筛选模块306，其中：In yet another optional embodiment, as shown in FIG. 4 , the apparatus may further include a judgment module 305 and a screening module 306, wherein:
判断模块305，用于在获取模块302获取特征检测模型输出的分析结果，作为胎儿超声图像的特征信息之后，根据胎儿超声图像的特征信息判断每个结构特征的轮廓是否满足确定出的预处理条件，当判断结果为否时，触发测量模块303执行上述的根据每个结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作。The judgment module 305 is used to obtain the analysis result output by the feature detection model in the acquisition module 302 as the feature information of the fetal ultrasound image, and judge whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image. , when the judgment result is no, triggering the measurement module 303 to perform the above-mentioned operation of measuring the contour of the structural feature according to the measurement method of each structural feature to obtain the geometric parameters of the structural feature.
筛选模块306，用于当判断模块305的判断结果为是时，从所有结构特征中筛选预处理条件的所有目标结构特征，并触发预处理模块304执行上述的基于确定出的预处理方式对每个结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的操作，所有结构特征包括所有目标结构特征。The screening module 306 is used to screen all the target structural features of the preprocessing condition from all the structural features when the judgment result of the judging module 305 is yes, and trigger the preprocessing module 304 to execute the above-mentioned preprocessing method based on the determined preprocessing method. Perform a preprocessing operation on the contour of each structural feature to obtain the contour of the preprocessed structural feature, and all structural features include all target structural features.
可见，实施图4所描述的装置还能够通过在获取到胎儿超声图像的特征信息之后，进一步判断胎儿超声图像的结构特征是否满足预处理条件，若不满足预处理条件，则直接执行结构特征的轮廓的测量操作，能够提高结构特征的几何参数的测量效率，从而提高胎儿的生长发育情况的确定效率；若满足预处理条件，对满足预处理条件的结构特征的轮廓执行预处理操作，能够获取到准确可靠的结构特征的轮廓，从而提高结构特征的几何参数的测量准确性以及可靠性，从而进一步提高胎儿的生长发育情况的确定准确性以及可靠性。It can be seen that, implementing the device described in FIG. 4 can further determine whether the structural features of the fetal ultrasound images satisfy the preprocessing conditions after acquiring the feature information of the fetal ultrasound images, and if the preprocessing conditions are not met, directly execute the structural features. The contour measurement operation can improve the measurement efficiency of the geometric parameters of the structural features, thereby improving the determination efficiency of the growth and development of the fetus; Accurate and reliable contours of structural features are obtained, thereby improving the measurement accuracy and reliability of geometric parameters of structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
在又一个可选的实施例中，如图4所示，判断模块305根据胎儿超声图像的特征信息判断每个结构特征的轮廓是否满足确定出的预处理条件的方式具体为：In yet another optional embodiment, as shown in FIG. 4 , the manner in which the judgment module 305 judges whether the contour of each structural feature satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image is specifically:
根据每个结构特征的类别判断该结构特征是否为确定出的需要执行预处理操作的结构特征的类别，当判断结果为是时，确定该结构特征满足确定出的预处理条件；或者，According to the category of each structural feature, it is determined whether the structural feature is the determined category of the structural feature that needs to perform a preprocessing operation, and when the judgment result is yes, it is determined that the structural feature satisfies the determined preprocessing condition; or,
根据每个结构特征的轮廓判断该结构特征的轮廓是否满足确定出的参数测量条件，当判断结果为否时，确定该结构特征满足确定出的预处理条件。According to the contour of each structural feature, it is determined whether the contour of the structural feature satisfies the determined parameter measurement condition, and when the judgment result is no, it is determined that the structural feature satisfies the determined preprocessing condition.
可见，实施图4所描述的装置还能够通过在判断出胎儿超声图像的结构特征的类别为需要执行预处理条件的结构特征的类别和/或该结构特征的轮廓不满足参数测量条件时，确定结构特征满足确定出的预处理条件，能够提高胎儿超声图像的结构特征需要执行预处理操作的确定准确性，还能够丰富胎儿超声图像的结构特征的轮廓需要执行预处理操作的确定方式，有利于获取到满足要求的结构特征，从而有利于进一步提高结构特征的几何参数的测量准确性以及可靠性，进而进一步提高胎儿的生长发育情况确定准确性以及可靠性。It can be seen that, implementing the device described in FIG. 4 can also be determined by determining that the category of the structural feature of the fetal ultrasound image is the category of the structural feature that needs to perform preprocessing conditions and/or the contour of the structural feature does not meet the parameter measurement conditions. The structural features satisfy the determined preprocessing conditions, which can improve the accuracy of determining that the structural features of the fetal ultrasound image need to perform preprocessing operations, and can also enrich the way that the contours of the structural features of the fetal ultrasound images need to perform preprocessing operations, which is beneficial to Obtaining the structural features that meet the requirements is beneficial to further improve the measurement accuracy and reliability of the geometric parameters of the structural features, thereby further improving the accuracy and reliability of determining the growth and development of the fetus.
实施例四Embodiment 4
请参阅图5，图5是本发明实施例公开的又一种胎儿结构特征的自动测量装置。其中，图5所描述的胎儿结构特征的自动测量装置可以应用于参数测量服务器中，其中，该参数测量服务器可以包括本地参数测量服务器或云参数测量服务器，本发明实施例不做限定。如图5所示，该胎儿结构特征的自动测量装置可以包括：Please refer to FIG. 5. FIG. 5 is another automatic measuring device for fetal structural features disclosed in an embodiment of the present invention. The automatic measurement device for fetal structural features described in FIG. 5 may be applied to a parameter measurement server, where the parameter measurement server may include a local parameter measurement server or a cloud parameter measurement server, which is not limited in this embodiment of the present invention. As shown in Figure 5, the automatic measuring device for fetal structural characteristics may include:
存储有可执行程序代码的存储器501；a memory 501 storing executable program code;
与存储器501耦合的处理器502；a processor 502 coupled to the memory 501;
进一步的，还可以包括与处理器502耦合的输入接口503以及输出接口504；Further, it may also include an input interface 503 coupled with the processor 502 and an output interface 504;
其中，处理器502调用存储器501中存储的可执行程序代码，用于执行实施例一或实施例二所描述的胎儿结构特征的自动测量方法中部分或者全部的步 骤。Wherein, the processor 502 invokes the executable program code stored in the memory 501 to execute some or all of the steps in the method for automatic measurement of fetal structural features described in Embodiment 1 or Embodiment 2.
实施例五Embodiment 5
本发明实施例公开了一种计算机可读存储介质，其存储用于电子数据交换的计算机程序，其中，该计算机程序使得计算机执行实施例一或实施例二所描述的胎儿结构特征的自动测量方法中部分或者全部的步骤。An embodiment of the present invention discloses a computer-readable storage medium, which stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the automatic measurement method for fetal structural features described in the first or second embodiment some or all of the steps.
实施例六Embodiment 6
本发明实施例公开了一种计算机程序产品，该计算机程序产品包括存储了计算机程序的非瞬时性计算机可读存储介质，且该计算机程序可操作来使计算机执行实施例一或实施例二所描述的胎儿结构特征的自动测量方法中部分或者全部的步骤。An embodiment of the present invention discloses a computer program product. The computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the description in the first embodiment or the second embodiment. Some or all of the steps in an automated method for measuring fetal structural features.
以上所描述的装置实施例仅是示意性的，其中所述作为分离部件说明的模块可以是或者也可以不是物理上分开的，作为模块显示的部件可以是或者也可以不是物理模块，即可以位于一个地方，或者也可以分布到多个网络模块上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in One place, or it can be distributed over multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.
通过以上的实施例的具体描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中,存储介质包括只读存储器(Read-Only Memory，ROM)、随机存储器(Random Access Memory，RAM)、可编程只读存储器(Programmable Read-only Memory，PROM)、可擦除可编程只读存储器(Erasable Programmable Read Only Memory，EPROM)、一次可编程只读存储器(One-time Programmable Read-Only Memory，OTPROM)、电子抹除式可复写只读存储器(Electrically-Erasable Programmable Read-Only Memory，EEPROM)、只读光盘(Compact Disc Read-OnlyMemory，CD-ROM)或其他光盘存储器、磁盘存储器、磁带存储器、或者能够用于携带或存储数据的计算机可读的任何其他介质。From the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation manner can be implemented by means of software plus a necessary general hardware platform, and certainly can also be implemented by means of hardware. Based on this understanding, the above-mentioned technical solutions can be embodied in the form of software products in essence or that make contributions to the prior art. The computer software products can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory. (Read-Only Memory, ROM), Random Access Memory (Random Access Memory, RAM), Programmable Read-only Memory (PROM), Erasable Programmable Read Only Memory (EPROM) ), One-time Programmable Read-Only Memory (OTPROM), Electronically Erasable Programmable Read-Only Memory (EEPROM), Compact Disc Read -Only Memory, CD-ROM) or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.
最后应说明的是：本发明实施例公开的一种胎儿结构特征的自动测量方法及装置所揭露的仅为本发明较佳实施例而已，仅用于说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解；其依然可以对前述各项实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或替换，并不使相应的技术方案的本质脱离本发明各项实施例技术方案的精神和范围。Finally, it should be noted that the method and device for automatic measurement of fetal structural features disclosed in the embodiments of the present invention are only preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, not to Limitation; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that; it is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements to some of the technical features. However, these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (10)

  1. 一种胎儿结构特征的自动测量方法，其特征在于，所述方法包括：An automatic measurement method for fetal structural characteristics, characterized in that the method comprises:
    将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析，并获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息，所述胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓；The acquired fetal ultrasound image is input into the determined feature detection model for analysis, and the analysis result output by the feature detection model is obtained as the feature information of the fetal ultrasound image, and the feature information of the fetal ultrasound image includes the a category of at least one structural feature of the fetal ultrasound image and a contour of at least one of the structural features;
    根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。According to the category of each of the structural features, a measurement method that matches the contour of the structural feature is determined, and a measurement operation is performed on the contour of the structural feature according to the measurement method of each of the structural features to obtain the geometric parameters of the structural feature. .
  2. 根据权利要求1所述的胎儿结构特征的自动测量方法，其特征在于，所述获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，所述方法还包括：The automatic measurement method for fetal structural features according to claim 1, wherein after acquiring the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, the method further comprises:
    基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，并触发执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作，每个所述结构特征的轮廓均为预处理后的结构特征的轮廓，所述预处理操作包括平滑处理操作。Perform a preprocessing operation on the contour of each of the structural features based on the determined preprocessing method, obtain the contour of the preprocessed structural feature, and trigger the execution of the measurement method according to each structural feature for the structure A measurement operation is performed on the contour of the feature to obtain the geometric parameters of the structural feature. The contour of each structural feature is the contour of the structural feature after preprocessing, and the preprocessing operation includes a smoothing operation.
  3. 根据权利要求2所述的胎儿结构特征的自动测量方法，其特征在于，所述获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息之后，所述方法还包括：The automatic measurement method for fetal structural features according to claim 2, wherein after obtaining the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, the method further comprises:
    根据所述胎儿超声图像的特征信息判断每个所述结构特征的轮廓是否满足确定出的预处理条件；Judging whether the contour of each of the structural features satisfies the determined preprocessing condition according to the feature information of the fetal ultrasound image;
    当判断结果为否时，触发执行所述的根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数的操作；When the judgment result is no, trigger the execution of the described operation of performing the measurement operation on the outline of the structural feature according to the measurement method of each of the structural features to obtain the geometric parameters of the structural feature;
    当判断结果为是时，从所有所述结构特征中筛选满足所述预处理条件的所有目标结构特征，并触发执行所述的基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓的操作，所有所述结构特征包括所有所述目标结构特征。When the judgment result is yes, screen all the target structural features that satisfy the preprocessing conditions from all the structural features, and trigger the execution of the outline of each structural feature based on the determined preprocessing method. The preprocessing operation is an operation of obtaining the contours of the preprocessed structural features, where all the structural features include all the target structural features.
  4. 根据权利要求2或3所述的胎儿结构特征的自动测量方法，其特征在于，每个所述结构特征的轮廓均对应多个节点；The automatic measuring method for fetal structural features according to claim 2 or 3, wherein the contour of each of the structural features corresponds to a plurality of nodes;
    以及，所述基于确定出的预处理方式对每个所述结构特征的轮廓执行预处理操作，得到预处理后的结构特征的轮廓，包括：And, performing a preprocessing operation on the contour of each of the structural features based on the determined preprocessing method, to obtain the contours of the preprocessed structural features, including:
    获取每个所述结构特征的轮廓对应的圆弧半径；Obtain the arc radius corresponding to the contour of each of the structural features;
    当每个所述结构特征的轮廓对应的圆弧半径大于等于确定出的圆弧半径阈值时，从该结构特征对应的所有所述节点中选取预设数量的目标节点，并按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述目标节点依次连接起来，得到预处理后的该结构特征的轮廓；When the arc radius corresponding to the outline of each of the structural features is greater than or equal to the determined arc radius threshold, select a preset number of target nodes from all the nodes corresponding to the structural feature, and select a preset number of target nodes according to each adjacent two All the target nodes corresponding to each of the structural features are sequentially connected in a manner of connecting each node to obtain the preprocessed outline of the structural feature;
    当每个所述结构特征的轮廓对应的圆弧半径不大于等于确定出的圆弧半径 阈值时，按照每相邻两个节点进行连接的方式将每个所述结构特征对应的所有所述节点依次连接起来，得到预处理后的该结构特征的轮廓。When the arc radius corresponding to the outline of each structural feature is not greater than or equal to the determined arc radius threshold, connect all the nodes corresponding to each structural feature in a manner of connecting every two adjacent nodes They are connected in sequence to obtain the preprocessed outline of the structural feature.
  5. 根据权利要求1-4任一项所述的胎儿结构特征的自动测量方法，其特征在于，所有所述结构特征包括颅脑结构特征、腹围结构特征以及其他结构特征中的至少一种，所述其他结构特征包括小脑结构特征、第三脑室结构特征、侧脑室结构特征、透明隔结构特征、肱骨结构特征以及股骨结构特征中的至少一种。The automatic measurement method for fetal structural features according to any one of claims 1-4, wherein all the structural features include at least one of cranial brain structural features, abdominal circumference structural features and other structural features, and the The other structural features include at least one of cerebellum structural features, third ventricle structural features, lateral ventricle structural features, septum pellucida structural features, humerus structural features, and femoral structural features.
  6. 根据权利要求5所述的胎儿结构特征的自动测量方法，其特征在于，所述根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数，包括：The automatic measurement method for fetal structural features according to claim 5, wherein the measurement method that matches the outline of each structural feature is determined according to the category of each structural feature, and the measurement method is determined according to each of the structural features. The feature measurement method performs a measurement operation on the outline of the structural feature to obtain the geometric parameters of the structural feature, including:
    当所述结构特征为所述颅脑结构特征时，基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数，所述颅脑结构特征的轮廓包括所述外轮廓和所述内轮廓；When the structural feature is the cranial structural feature, the geometric parameters corresponding to the cranial structural feature are determined based on the inner contour of the cranial structural feature and the outer contour of the cranial structural feature. The contour of the brain structure feature includes the outer contour and the inner contour;
    当所述结构特征为所述腹围结构特征时，获取所述腹围结构特征的轮廓对应的周长，作为所述腹围结构特征的几何参数；When the structural feature is the abdominal circumference structural feature, acquiring the perimeter corresponding to the contour of the abdominal circumference structural feature as a geometric parameter of the abdominal circumference structural feature;
    当所述结构特征为所述其他结构特征时，基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，所述图形拟合方式包括最小外接矩形拟合方式和/或椭圆拟合方式。When the structural feature is the other structural feature, fit the contour of each other structural feature based on the determined graph fitting method, obtain the target node of the other structural feature, and based on each of the other structural features The target node of the feature determines the geometric parameters of the other structural feature, and the graph fitting method includes a minimum circumscribed rectangle fitting method and/or an ellipse fitting method.
  7. 根据权利要求6所述的胎儿结构特征的自动测量方法，其特征在于，所述基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，包括：The automatic measurement method for fetal structural features according to claim 6, wherein the contour of each of the other structural features is fitted based on the determined graphic fitting method to obtain the target node of the other structural features, And based on the target node of each of the other structural features, determine the geometric parameters of the other structural features, including:
    当所述图形拟合方式为所述最小外接矩形拟合方式时，基于确定出的外接矩形拟合所述其他结构特征的轮廓，当所述外接矩形的其中一组对边均与所述其他结构特征的轮廓存在目标交点时，停止拟合所述其他结构特征的轮廓，得到所述其他结构特征的轮廓的最小外接矩形，并根据所述其他结构特征的类别确定所述其他结构特征的扫查方向；When the graph fitting mode is the minimum circumscribed rectangle fitting mode, the contours of the other structural features are fitted based on the determined circumscribed rectangle. When the contour of the structural feature has a target intersection, stop fitting the contour of the other structural feature, obtain the minimum circumscribed rectangle of the contour of the other structural feature, and determine the sweep of the other structural feature according to the category of the other structural feature. check the direction;
    基于所述扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条扫查线段，并从所有所述扫查线段中选择距离值最大的扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the scanning direction, obtain multiple scanning line segments of the contours of the other structural features, and select the scanning line segment with the largest distance value from all the scanning line segments Line segments, as geometric parameters of the other structural features;
    其中，所述扫查方向包括第一扫查方向或第二扫查方向中的其中一种，其中，所述第一扫查方向用于表示从平行于所述最小外接矩形的短边对所述其他结构特征的轮廓进行扫查，所述第二扫查方向用于表示从平行于所述最小外接矩形的长边对所述其他结构特征的轮廓进行扫查。Wherein, the scanning direction includes one of a first scanning direction or a second scanning direction, wherein the first scanning direction is used to indicate that the scanning direction is parallel to the short side of the minimum circumscribed rectangle. The contour of the other structural feature is scanned, and the second scanning direction is used to indicate that the contour of the other structural feature is scanned from the long side parallel to the minimum circumscribed rectangle.
  8. 根据权利要求6所述的胎儿结构特征的自动测量方法，其特征在于，所述基于确定出的图形拟合方式拟合每个所述其他结构特征的轮廓，得到该其他 结构特征的目标节点，并基于每个所述其他结构特征的目标节点，确定该其他结构特征的几何参数，包括：The automatic measurement method for fetal structural features according to claim 6, wherein the contour of each of the other structural features is fitted based on the determined graphic fitting method to obtain the target node of the other structural features, And based on the target node of each of the other structural features, determine the geometric parameters of the other structural features, including:
    当所述图形拟合方式为所述椭圆拟合方式时，基于确定出的椭圆拟合算法拟合所述其他结构特征的轮廓，得到拟合后的所述其他结构特征的轮廓，拟合后的所述其他结构特征的轮廓的形状为椭圆形；When the graph fitting method is the ellipse fitting method, the contours of the other structural features are fitted based on the determined ellipse fitting algorithm, and the fitted contours of the other structural features are obtained. The shape of the outline of the other structural features is an ellipse;
    根据所述其他结构特征的类别确定所述其他结构特征的目标扫查方向；Determine the target scanning direction of the other structural features according to the categories of the other structural features;
    基于所述目标扫查方向对所述其他结构特征的轮廓执行扫查操作，得到所述其他结构特征的轮廓的多条目标扫查线段，并从所有所述目标扫查线段中选择距离值最大的目标扫查线段，作为所述其他结构特征的几何参数；Perform a scanning operation on the contours of the other structural features based on the target scanning direction, obtain multiple target scanning line segments of the contours of the other structural features, and select the largest distance value from all the target scanning line segments The target scanning line segment is used as the geometric parameter of the other structural features;
    其中，所述目标扫查方向包括第一目标扫查方向或第二目标扫查方向中的其中一种，其中，所述第一目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的短轴对所述其他结构特征的轮廓进行扫查，所述第二目标扫查方向用于表示从平行于拟合后的所述其他结构特征的轮廓的长轴对所述其他结构特征的轮廓进行扫查。Wherein, the target scanning direction includes one of a first target scanning direction or a second target scanning direction, wherein the first target scanning direction is used to represent a direction from parallel to the fitted The short axis of the contour of the other structural feature scans the contour of the other structural feature, and the second target scanning direction is used to indicate that the long axis of the contour of the other structural feature is parallel to the fitting. Scan the contours of other structural features described above.
  9. 根据权利要求6-8任一项所述的胎儿结构特征的自动测量方法，其特征在于，所述基于所述颅脑结构特征的内轮廓与所述颅脑结构特征的外轮廓，确定所述颅脑结构特征对应的几何参数，包括：The automatic measurement method for fetal structural features according to any one of claims 6-8, wherein the determination of the The geometric parameters corresponding to the structural features of the brain include:
    获取所述颅脑结构特征的内轮廓的第一周长与所述颅脑结构特征的外轮廓的第二周长，并基于所述第一周长和所述第二周长，确定所述颅脑结构特征对应的头围几何参数；Obtain the first circumference of the inner contour of the cranial structure feature and the second circumference of the outer contour of the skull structure feature, and determine the first circumference based on the first circumference and the second circumference The geometric parameters of the head circumference corresponding to the structural features of the brain;
    确定所述颅脑结构特征的脑中线对应的中垂线与所述颅脑结构特征的外轮廓的第一交点以及所述中垂线与所述颅脑结构特征的内轮廓的第二交点，并基于所述第一交点与所述第二交点，确定所述颅脑结构特征对应的双顶径几何参数。determining the first intersection point of the midline corresponding to the midline of the cranial structure feature and the outer contour of the craniocerebral structure feature and the second intersection point of the perpendicular line and the inner contour of the craniocerebral structure feature, And based on the first intersection point and the second intersection point, the biparietal diameter geometric parameter corresponding to the cranial brain structural feature is determined.
  10. 一种胎儿结构特征的自动测量装置，其特征在于，所述装置包括：An automatic measuring device for fetal structural characteristics, characterized in that the device comprises:
    分析模块，将获取到的胎儿超声图像输入确定出的特征检测模型中进行分析；An analysis module, which inputs the acquired fetal ultrasound image into the determined feature detection model for analysis;
    获取模块，用于获取所述特征检测模型输出的分析结果，作为所述胎儿超声图像的特征信息，所述胎儿超声图像的特征信息包括该胎儿超声图像的至少一个结构特征的类别以及至少一个该结构特征的轮廓；The acquisition module is configured to acquire the analysis result output by the feature detection model as the feature information of the fetal ultrasound image, where the feature information of the fetal ultrasound image includes the category of at least one structural feature of the fetal ultrasound image and at least one of the features of the fetal ultrasound image. outline of structural features;
    测量模块，用于根据每个所述结构特征的类别确定与该结构特征的轮廓相匹配的测量方式，并根据每个所述结构特征的测量方式对该结构特征的轮廓执行测量操作，得到该结构特征的几何参数。The measurement module is configured to determine a measurement mode that matches the contour of the structural feature according to the category of each of the structural features, and perform a measurement operation on the contour of the structural feature according to the measurement mode of each of the structural features to obtain the Geometric parameters of structural features.
PCT/CN2021/096822 2020-09-24 2021-05-28 Automatic measurement method and device for fetal structural characteristic WO2022062459A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011014990.6 2020-09-24
CN202011014990.6A CN112233167B (en) 2020-09-24 2020-09-24 Automatic measurement method and device for fetal structural characteristics

Publications (1)

Publication Number Publication Date
WO2022062459A1 true WO2022062459A1 (en) 2022-03-31

Family

ID=74108375

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/096822 WO2022062459A1 (en) 2020-09-24 2021-05-28 Automatic measurement method and device for fetal structural characteristic

Country Status (2)

Country Link
CN (1) CN112233167B (en)
WO (1) WO2022062459A1 (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112233167B (en) * 2020-09-24 2024-06-21 广州爱孕记信息科技有限公司 Automatic measurement method and device for fetal structural characteristics
CN115482190A (en) * 2021-11-10 2022-12-16 中山大学附属第七医院(深圳) Fetal heart structure segmentation measurement method and device and computer storage medium
CN118049938A (en) * 2024-04-15 2024-05-17 西安第六镜网络科技有限公司 Rail measuring method, device, electronic equipment, readable storage medium and system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103239249A (en) * 2013-04-19 2013-08-14 深圳大学 Measurement method for fetus ultrasound image
US20140185895A1 (en) * 2012-12-31 2014-07-03 General Electric Company Methods and systems for automated soft tissue segmentation, circumference estimation and plane guidance in fetal abdominal ultrasound images
CN110680399A (en) * 2019-10-25 2020-01-14 深圳度影医疗科技有限公司 Automatic measurement method of prenatal ultrasound image, storage medium and ultrasound equipment
CN110742654A (en) * 2019-11-05 2020-02-04 深圳度影医疗科技有限公司 Method for positioning and measuring standard tangent plane based on three-dimensional ultrasonic image
CN112233167A (en) * 2020-09-24 2021-01-15 广州爱孕记信息科技有限公司 Automatic measurement method and device for structural characteristics of fetus

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110613480B (en) * 2019-01-14 2022-04-26 广州爱孕记信息科技有限公司 Fetus ultrasonic dynamic image detection method and system based on deep learning

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140185895A1 (en) * 2012-12-31 2014-07-03 General Electric Company Methods and systems for automated soft tissue segmentation, circumference estimation and plane guidance in fetal abdominal ultrasound images
CN103239249A (en) * 2013-04-19 2013-08-14 深圳大学 Measurement method for fetus ultrasound image
CN110680399A (en) * 2019-10-25 2020-01-14 深圳度影医疗科技有限公司 Automatic measurement method of prenatal ultrasound image, storage medium and ultrasound equipment
CN110742654A (en) * 2019-11-05 2020-02-04 深圳度影医疗科技有限公司 Method for positioning and measuring standard tangent plane based on three-dimensional ultrasonic image
CN112233167A (en) * 2020-09-24 2021-01-15 广州爱孕记信息科技有限公司 Automatic measurement method and device for structural characteristics of fetus

Also Published As

Publication number Publication date
CN112233167B (en) 2024-06-21
CN112233167A (en) 2021-01-15

Similar Documents

Publication Publication Date Title
WO2022062459A1 (en) Automatic measurement method and device for fetal structural characteristic
EP3770850B1 (en) Medical image identifying method, model training method, and computer device
US9208582B2 (en) Image analyzing system and method
US9848765B2 (en) Quantifying a blood vessel reflection parameter of the retina
WO2022095895A1 (en) Vascular stenosis analysis method and apparatus
WO2021208739A1 (en) Method and apparatus for evaluating blood vessel in fundus color image, and computer device and medium
CN111462049B (en) Automatic lesion area form labeling method in mammary gland ultrasonic radiography video
Rahmatullah et al. Quality control of fetal ultrasound images: Detection of abdomen anatomical landmarks using adaboost
US20240074658A1 (en) Method and system for measuring lesion features of hypertensive retinopathy
WO2022062461A1 (en) Fetal section determination method and device based on ultrasonic dynamic image
AU2012207076A1 (en) Automated determination of arteriovenous ratio in images of blood vessels
WO2022062460A1 (en) Method and device for determining imaging quality control of fetal ultrasound image
WO2022048171A1 (en) Method and apparatus for measuring blood vessel diameter in fundus image
Zhu et al. Automatic measurement of fetal femur length in ultrasound images: a comparison of random forest regression model and SegNet
Koprowski et al. Assessment of significance of features acquired from thyroid ultrasonograms in Hashimoto's disease
CN112348780A (en) Fetal heart measuring method and device
CN111275755A (en) Mitral valve orifice area detection method, system and equipment based on artificial intelligence
WO2020215485A1 (en) Fetal growth parameter measurement method, system, and ultrasound device
CN114898882A (en) Method and system for ultrasound-based assessment of right heart function
CN117547306B (en) Left ventricular ejection fraction measurement method, system and device based on M-type ultrasound
Yasrab et al. End-to-end first trimester fetal ultrasound video automated crl and nt segmentation
US20220061920A1 (en) Systems and methods for measuring the apposition and coverage status of coronary stents
CN107169978B (en) Ultrasonic image edge detection method and system
CN116912229A (en) Ultrasonic standard section detection method for cross section of lateral ventricle of fetus
CN115762787B (en) Eyelid disease operation curative effect evaluation method and system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21870846

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21870846

Country of ref document: EP

Kind code of ref document: A1