CN115063395A - Ultrasonic image processing method, device, equipment and medium - Google Patents
Ultrasonic image processing method, device, equipment and medium Download PDFInfo
- Publication number
- CN115063395A CN115063395A CN202210801337.7A CN202210801337A CN115063395A CN 115063395 A CN115063395 A CN 115063395A CN 202210801337 A CN202210801337 A CN 202210801337A CN 115063395 A CN115063395 A CN 115063395A
- Authority
- CN
- China
- Prior art keywords
- contour
- standard
- determining
- cerebellar
- standard section
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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
- G06T2207/30016—Brain
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The application discloses an ultrasonic image processing method, an ultrasonic image processing device, ultrasonic image processing equipment and an ultrasonic image processing medium, wherein the ultrasonic image processing method comprises the following steps: identifying a standard fetal head section in an ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode; inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises a contour of a key structure in the fetal head and a brain midline; determining a measurement of the fetal head standard section based on the critical structure contour and the brain midline. The accuracy of the standard section measurement of the fetal head can be improved.
Description
Technical Field
The present application relates to the field of ultrasound image processing technologies, and in particular, to an ultrasound image processing method, apparatus, device, and medium.
Background
During the ultrasonic scanning of the fetus, the standard section of the head of the fetus is required to be measured to obtain the ultrasonic scanning result. At present, two measurement modes mainly exist, one mode is to manually measure the standard section of the head of the fetus, but the mode depends on the experience of a doctor and takes time for correction; the other method is to use a deep learning technology for measurement, but when the head key structure profile is irregular, deviation is easy to occur, and the measurement result is inaccurate.
Disclosure of Invention
In view of this, an object of the present application is to provide an ultrasound image processing method, apparatus, device and medium, which can improve the accuracy of the measurement of the standard section of the fetal head. The specific scheme is as follows:
in a first aspect, the present application discloses an ultrasound image processing method, including:
identifying a standard fetal head section in an ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode;
inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises a contour of a key structure in the fetal head and a brain midline;
determining a measurement of the fetal head standard section based on the critical structure contour and the brain midline.
Optionally, the fetal head standard section comprises a thalamus standard section and a cerebellum standard section, the key structure contour comprises a target thalamus contour and a target cerebellum contour, and the brain midline comprises a first brain midline and a second brain midline; correspondingly, the measurement result of determining the standard section of the fetal head based on the contour of the key structure and the brain midline comprises
Determining a thalamic double apical diameter measurement of the thalamic standard section based on a target thalamic contour of the thalamic standard section and the first brain centerline;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the target cerebellar contour of the cerebellar standard section and the second brain midline.
Optionally, the determining the thalamic double apical diameter measurement of the thalamic standard cut based on the target thalamic contour of the thalamic standard cut and the first brain midline comprises:
determining a slope of a dual apical diameter line based on the first cerebral midline;
determining a double-top-diameter straight line based on the slope of the double-top-diameter straight line;
determining a double apical diameter line segment based on the double apical diameter line and the target thalamic contour;
determining thalamic double apical diameter measurements of the thalamic standard section based on the double apical diameter line segments.
Optionally, the double-crest-diameter straight line is determined based on the slope of the double-crest-diameter straight line; determining a double apical diameter line segment based on the double apical diameter line and the target thalamic contour, including:
determining a centroid of the target thalamic contour and determining a target pixel range based on the centroid;
determining a plurality of double-top-diameter straight lines based on the target pixel range and the slope of the double-top-diameter straight lines;
determining a plurality of line segments based on the plurality of double apical diameter lines and the target thalamic contour;
and determining the longest line segment in the line segments as a double-top-diameter line segment.
Optionally, the determining a thalamic double apical diameter measurement of the thalamic standard section based on the double apical diameter line segment includes:
determining a lower bone ring contour from the thalamus standard tangent plane, and determining the lowest point of the lower bone ring contour as the final bottom end point of the double-top diameter line segment to obtain a final double-top diameter line segment;
determining a thalamic double apical diameter measurement of the thalamic standard tangent plane based on the final double apical diameter line segment.
Optionally, said determining the contour of the inferior bone ring from said thalamic standard cut plane comprises:
determining a first area range based on the bottom end point of the double-top-diameter line segment;
intercepting a first region image from the thalamic standard section based on the first region range;
an inferior bone ring contour is determined from the first region image.
Optionally, the determining a cerebellar transverse diameter measurement result of the cerebellar standard cut plane based on the target cerebellar contour of the cerebellar standard cut plane and the second brain centerline includes:
determining a slope of a cerebellar transverse diameter line based on the second cerebral midline;
determining a cerebellar transverse diameter line based on the slope of the cerebellar transverse diameter line;
determining a cerebellar transverse diameter line segment based on the cerebellar transverse diameter straight line and the target cerebellar contour;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the cerebellar transverse diameter line segment.
Optionally, the determining a cerebellum transverse diameter measurement result of the cerebellum standard tangent plane based on the cerebellum transverse diameter line segment includes:
determining a second area range and a third area range respectively based on two end points of the cerebellar transverse diameter line segment;
respectively intercepting a second area image and a third area image from the cerebellum standard tangent plane based on the second area range and the third area range;
respectively determining cerebellum local contours in the second area image and the third area image;
determining the intersection point of the cerebellum transverse diameter straight line and the cerebellum local contour;
determining a final cerebellar transverse diameter line segment based on the intersection points;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the final cerebellar transverse diameter line segment.
Optionally, the fetal head standard section further includes a lateral brain standard section, and the key structure contour further includes a target lateral ventricle contour; correspondingly, the method further comprises the following steps:
and determining the inner diameter measurement result of the lateral ventricle posterior horn of the lateral brain standard section based on the target lateral ventricle contour of the lateral brain standard section.
Optionally, the determining a lateral ventricle posterior angle inner diameter measurement result of the lateral brain standard cut plane based on the target lateral ventricle contour of the lateral brain standard cut plane includes:
determining location information of a choroid plexus in a lateral ventricle based on the target lateral ventricle contour;
determining a lateral ventricles posterior horn inner diameter line based on the position information;
determining a lateral ventriculo-caudal posterior radius line segment based on the lateral ventriculo-medial diameter line and the target lateral ventriculo-ventricular profile;
and determining the lateral ventricle posterior angle inner diameter measurement result of the lateral ventricle posterior angle standard tangent plane based on the lateral ventricle posterior angle inner diameter line segment.
Optionally, the determining position information of the choroid plexus in the lateral ventricle based on the target lateral ventricle contour includes:
determining an outgoing lateral ventricle image from the lateral brain standard slice based on the target lateral ventricle contour;
multiplying the mask of the target lateral ventricle outline and the local lateral ventricle image to obtain a target lateral ventricle image;
performing thinning processing on the mask of the target lateral ventricle outline to obtain a first central axis;
multiplying the first central axis by the target lateral ventricle image to obtain a second central axis of the choroid plexus;
respectively performing straight line fitting on the first central axis and the second central axis to obtain a first line segment and a second line segment;
determining location information of the choroid plexus in a lateral ventricle based on the first and second line segments.
Optionally, the training process of the image segmentation model includes:
acquiring a first training sample set; the first training sample set comprises a fetal head standard section sample and label information corresponding to the fetal head standard section sample; the fetal head standard section sample comprises a thalamus standard section sample, a cerebellum standard section sample and a lateral brain standard section sample; the label information of the thalamus standard section sample comprises thalamus contour lines and brain midline, the label information of the cerebellum standard section sample comprises cerebellum contour lines and brain midline, and the label information of the lateral cerebrum standard section comprises lateral ventricle contour lines;
inputting the fetal head standard section sample into a first initial model to obtain a thalamus contour, a cerebellum contour, a brain midline and a lateral ventricle contour corresponding to the fetal head standard section sample;
calculating a training loss based on the thalamic contour, the cerebellar contour, the brain midline, the lateral ventricle contour, and label information of the fetal head standard sectional sample, and performing parameter adjustment on the first initial model based on the training loss;
and if the condition that the training completion condition is met is detected, determining the first initial model after parameter adjustment as an image segmentation model.
Optionally, the identifying a standard fetal head section in an ultrasound video, or identifying a standard fetal head section in an ultrasound real-time mode includes:
identifying a standard fetal head section in an ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode by using a standard section identification model;
the standard section recognition model is obtained by training a second initial model by using a second training sample set, wherein the second training sample set comprises a fetal head standard section sample, a non-fetal head standard section sample and label information.
In a second aspect, the present application discloses an ultrasound image processing apparatus, comprising:
the standard section module is used for identifying a standard section of the head of the fetus in an ultrasonic video or identifying the standard section of the head of the fetus in an ultrasonic real-time mode;
the image segmentation module is used for inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises a contour of a key structure in the fetal head and a brain midline;
and the image measuring module is used for determining the measurement result of the standard fetal head section based on the key structure contour and the brain midline.
In a third aspect, the present application discloses an ultrasound device comprising a processor and a memory; wherein the content of the first and second substances,
the memory is used for storing a computer program;
the processor is used for executing the computer program to realize the ultrasonic image processing method.
In a fourth aspect, the present application discloses a computer readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the ultrasound image processing method as described above.
Therefore, the standard fetal head section in the ultrasonic video is identified, then the standard fetal head section is input into an image segmentation model, the segmentation result of the standard fetal head section is obtained and comprises the contour of the key structure in the fetal head and the brain midline, and finally the measurement result of the standard fetal head section is determined based on the contour of the key structure and the brain midline. That is, according to the method, the image segmentation model is firstly utilized to determine the outline of the key structure in the head of the fetus and the brain midline, then the brain midline is utilized to assist automatic measurement, the direction accuracy of a measurement item in the standard section of the head of the fetus is guaranteed, and the accuracy of the measurement of the standard section of the head of the fetus can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a flow chart of a method for processing ultrasound images as disclosed herein;
FIG. 2 is a schematic diagram of a thalamus standard section labeling and segmentation disclosed in the present application;
FIG. 3 is a schematic diagram of a cerebellum standard section marking and segmentation disclosed in the present application;
FIG. 4 is a schematic diagram of a standard section marking and segmentation of a lateral brain disclosed in the present application;
FIG. 5 is a schematic view of an exemplary ultrasound image processing system disclosed herein;
fig. 6 is a schematic structural diagram of an ultrasound image processing apparatus according to the present disclosure;
FIG. 7 is a block diagram of an ultrasound device as disclosed herein.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
At present, two measurement modes mainly exist for a standard fetal head section, one mode is to manually measure the standard fetal head section, but the mode depends on the experience of a doctor and takes time for correction; the other method is to use a deep learning technology for measurement, but when the head key structure profile is irregular, deviation is easy to occur, and the measurement result is inaccurate. Therefore, the ultrasonic image processing scheme is provided, and the accuracy of the measurement of the standard section of the head of the fetus can be improved.
Referring to fig. 1, an embodiment of the present application discloses an ultrasound image processing method, including:
step S11: and identifying a standard fetal head section in the ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode.
In one embodiment, the standard section identification model can be used for identifying the standard section of the fetal head in an ultrasonic video or identifying the standard section of the fetal head in an ultrasonic real-time mode; the standard section recognition model is obtained by training a second initial model by using a second training sample set, wherein the second training sample set comprises a fetal head standard section sample, a non-fetal head standard section sample and label information.
Further, the training process of the standard tangent plane recognition model may include: and carrying out thalamus, cerebellum and lateral brain standard section labeling on the collected ultrasonic video on labeling software. The first initial model may use the target detection algorithm RCNN (i.e., Region Convolutional Neural Networks) series, YOLO series, RetinaNet, etc. And inputting the marked ultrasonic image and the label thereof into the first initial model for training, and performing iterative optimization to finally train a model with the highest recognition accuracy.
It should be noted that, in the ultrasonic scanning of the fetus, the thalamic, cerebellar and lateral cerebral sections of the fetal head are very basic and important sections for diagnosing whether the fetal head is normally developed. The judgment standard of the thalamus standard section is as follows: the skull is in an elliptical strong echogenic ring, the cerebral hemispheres on two sides are symmetrical, the midline of the brain is in the middle, and a transparent separate cavity, the thalamus on two sides are symmetrical and a fissure-like third ventricle between the thalamus is clearly shown. The measurement item is the double apical diameter BPD, also called fetal head double apical diameter, which is the length of the widest part between the left side and the right side of the fetal head, also called head large transverse diameter. The judgment standard of the cerebellum standard section is as follows: the skull is an elliptical strong echo ring, a clear cerebellar hemisphere, a transparent partition cavity which is symmetrical left and right and in front. The measurement term of the cerebellar transverse diameter refers to the horizontal length of the cerebellum of a fetus, is usually used for predicting the gestational week and is an ideal index reflecting the actual gestational week. The standard judgment standard of the lateral cerebral section is as follows: the lateral ventricles posterior horn appeared clear, anechoic, with a highly echogenic choroid plexus inside, but not completely filling the posterior horn. The center of the image still shows the bilateral thalamus, with the midline of the brain visible. The lateral ventricles are in the lateral aspect of the frontal angle almost parallel to the sickle of the brain, and the occipital angle is laterally spaced further from the midline of the brain. The inside diameter measuring vernier for measuring the posteroventricular diameter is arranged on the inner side edge of the widest part of the lateral ventricles and is perpendicular to the long axis of the lateral ventricles, the maximum width of the vernier is measured, whether ventricular dilatation and hydrocephalus exist or not can be judged, and the inside diameter of the occipital horn of the lateral ventricles of the fetus is smaller than 10mm in the whole gestation period.
Therefore, a doctor does not need to manually obtain the standard fetal head section, the standard fetal head section is automatically obtained in the scanning process, and the diagnosis efficiency is improved.
Step S12: inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises the contour of a key structure in the fetal head and the brain midline.
Wherein the fetal head standard section comprises a thalamus standard section and a cerebellum standard section, the key structure contour comprises a target thalamus contour and a target cerebellum contour, and the brain midline comprises a first brain midline and a second brain midline; in addition, in one embodiment, the fetal head standard section further comprises a lateral brain standard section, and the critical structure contour further comprises a target lateral ventricle contour.
In one embodiment, the training process of the image segmentation model may include:
step 00: acquiring a first training sample set; the first training sample set comprises a fetal head standard section sample and label information corresponding to the fetal head standard section sample; the fetal head standard section sample comprises a thalamus standard section sample, a cerebellum standard section sample and a lateral brain standard section sample; the label information of the thalamus standard section sample comprises thalamus contour lines and brain midline, the label information of the cerebellum standard section sample comprises cerebellum contour lines and brain midline, and the label information of the lateral cerebrum standard section comprises lateral ventricle contour lines.
Step 01: inputting the fetal head standard section sample into a first initial model to obtain a thalamus contour, a cerebellum contour, a brain midline and a lateral ventricle contour corresponding to the fetal head standard section sample;
step 02: calculating a training loss based on the thalamic contour, the cerebellar contour, the midline of the brain, the lateral ventricle contour, and label information of the fetal head standard sectional sample, and performing parameter adjustment on the first initial model based on the training loss.
Step 03: and if the condition that the training completion condition is met is detected, determining the first initial model after parameter adjustment as an image segmentation model.
The training completion condition may be that the training loss is less than a preset loss threshold, and the iteration number reaches a preset number threshold.
For example, referring to fig. 2, fig. 3 and fig. 4, fig. 2 is a schematic diagram illustrating a thalamic standard section labeling and segmentation disclosed in the present application; FIG. 3 is a schematic diagram of a cerebellum standard section marking and segmentation disclosed in the present application; fig. 4 is a schematic diagram of a standard section marking and segmentation of a lateral brain disclosed in the present application. The collected thalamus, cerebellum and lateral brain standard sections can be labeled on labeling software, as shown in fig. 2(a), a thalamus standard section labeling schematic diagram is shown, the thalamus standard section labeling schematic diagram comprises three labels of thalamus-craniocerebral contour lines, namely thalamus contour lines, brain midline lines and thalamus-double apical diameter measuring lines, as shown in fig. 3(a), a cerebellum standard section labeling schematic diagram is shown, the cerebellum standard section labeling schematic diagram comprises three labels of cerebellum hemisphere contour lines, namely cerebellum contour lines, brain midline lines and cerebellum transverse diameter measuring lines, as shown in fig. 4(a), a lateral brain standard section labeling schematic diagram comprises two labels of lateral ventriculo-medial border contour lines, namely lateral ventriculum contour lines and lateral ventriculum posterior corner inner diameter measuring lines, and the labeled images and the labels are input to a second initial model for training. In one embodiment, the second initial model may be a model improved based on a human pose keypoint detection model, the original keypoint detection model is used to detect a plurality of keypoints of a human body, including a nose, eyes, ears, shoulders, elbows, wrists, knees, and the like, in the embodiment of the present application, the original plurality of channels are changed into 4 channels, the first three channels are changed into semantic segmentation channels for learning thalamic contours, cerebellar contours, and lateral ventricle contours, and the fourth channel is changed into a line detection channel for learning midline, so as to complete multitask measurement for simultaneously outputting fetal thalamic contours, cerebellums, lateral brain contour masks, and midline. That is, the model has 4 output channels, which respectively output the inference results of the thalamic contour, the brain midline, the cerebellar contour and the lateral ventricle contour, the inference results are masks of the thalamic contour, the brain midline, the cerebellar contour and the lateral ventricle contour, contour extraction is performed on the masks of the thalamic contour, the cerebellar contour and the lateral ventricle contour to obtain the thalamic contour, the cerebellar contour and the lateral ventricle contour, straight line fitting is performed on the masks of the brain midline to obtain the brain midline, and the brain midline is drawn on an image, as shown in fig. 2(b), fig. 2(b) is a segmentation result schematic diagram of a standard thalamic section, which comprises the thalamic contour and the brain midline; FIG. 3(b) is a diagram showing the result of segmentation of a standard section of cerebellum, including the contour of the cerebellum and the midline of the brain; FIG. 4(b) is a diagram showing the segmentation result of the standard lateral cerebral slice, including the lateral ventricle contour. And (5) performing iterative training optimization, and finally training a model with the highest segmentation accuracy. In another embodiment, the second initial model may employ MobileNet.
It should be noted that the thalamus-double-vertex diameter measurement line, cerebellum transverse diameter measurement line, and lateral ventricle posterior angle inner diameter measurement line in the label do not participate in model training and are used for testing the accuracy of measurement, for example, in the test stage, the coincidence degree of the thalamus-double-vertex diameter measurement line, the cerebellum transverse diameter measurement line, and the lateral ventricle posterior angle inner diameter measurement line with the thalamus double-vertex diameter line segment, the cerebellum transverse diameter line segment, and the lateral ventricle posterior angle inner diameter line segment determined based on the segmentation result can be respectively calculated to measure the accuracy of measurement.
Step S13: determining a measurement of the fetal head standard section based on the critical structure contour and the brain midline.
Wherein, step S13 may specifically include:
step 10: determining a thalamic double apical diameter measurement of the thalamic standard cut based on a target thalamic contour of the thalamic standard cut and the first brain midline.
In one embodiment, step 10 may specifically include the following steps:
step 100: determining a slope of a dual apical diameter line based on the first cerebral midline.
It will be appreciated that the first cerebral midline and the double apical diameter line are in a perpendicular relationship, and the slope of the double apical diameter line can be determined based on the perpendicular relationship.
Step 101: and determining the double-top-diameter straight line based on the slope of the double-top-diameter straight line.
Step 102: determining a double apical diameter line segment based on the double apical diameter line and the target thalamic contour.
In one embodiment, step 102 may specifically include:
step 1020: a centroid of the target thalamic contour is determined, and a target pixel range is determined based on the centroid. For example, a region of 10 pixels left and right of the centroid can be taken with the centroid as the center to obtain the target pixel range.
Step 1021: and determining a plurality of double-top-diameter straight lines based on the target pixel range and the slope of the double-top-diameter straight lines. Namely, a plurality of double-top-diameter straight lines are determined according to the slope of the double-top-diameter straight lines after passing through each pixel point in the target pixel range.
Step 1021: determining a plurality of line segments based on the plurality of double apical diameter lines and the target thalamic contour.
In one embodiment, the intersection points of the multiple double apical diameter lines and the target thalamic contour may be determined separately, and the two intersection points corresponding to each double apical diameter line may be connected to obtain multiple line segments. In another embodiment, the intersection of the plurality of double top diameter lines and the mask of the target thalamic contour may be determined separately, resulting in a plurality of line segments.
Step 1022: and determining the longest line segment in the line segments as a double-top-diameter line segment.
It should be pointed out that, compare with the scheme of traditional ellipse fitting, calculate the perpendicular line as the direction of double apical radius according to the brain central line, measurement accuracy is higher, avoids the ellipse deviation of fitting out when thalamus profile is irregular great, leads to the problem that the direction of double apical radius appears great skew.
Step 103: determining thalamic double apical diameter measurements of the thalamic standard section based on the double apical diameter line segments.
Further, step 103 may specifically include:
step 1030: and determining a lower bone ring contour from the thalamus standard section, and determining the lowest point of the lower bone ring contour as the final bottom end point of the double-top-diameter line segment to obtain the final double-top-diameter line segment.
In one embodiment, a first region range may be determined based on a bottom endpoint of the double top diameter line segment; intercepting a first region image from the thalamic standard section based on the first region range; an inferior bone ring contour is determined from the first region image. And if the lower bone ring contour cannot be determined from the first region image, moving the bottom end points of the double-top-diameter line segments upwards according to a preset upwards moving value to obtain the final bottom end points of the double-top-diameter line segments.
For example, a square area is determined by taking the bottom end point of the double-top-diameter line segment as the center and taking the first preset size as the first area range. And taking out the square area image of the corresponding position of the thalamic standard tangent plane, carrying out threshold segmentation, and taking the maximum contour as the contour of the lower bone ring. If the lower bone ring contour cannot be found, the bottom end points of the double-top-diameter line pairs are moved up by 10 pixels to obtain final double-top-diameter line segments, and the final double-top-diameter line segments are drawn on the image, as shown in fig. 2 (b). It should be noted that, because the bottom end of the double-top-diameter line segment cannot pass through the inferior bone ring according to clinical requirements, the embodiment of the application can ensure that the bottom end of the double-top-diameter line segment cannot pass through the inferior bone ring.
Step 1031: determining a thalamic double apical diameter measurement of the thalamic standard section based on the final double apical diameter line segment.
The method and the device can determine the length of the final double-apical-diameter line segment as the thalamus double-apical-diameter measurement result of the thalamus standard tangent plane. Further, the double-top-diameter measurement result is displayed on an ultrasonic equipment interface. It should be noted that, before 31 weeks of pregnancy, the growth rate is 3 mm/week, after 31-36 weeks of pregnancy, the growth rate is 1.5 mm/week, and after 36 weeks of pregnancy, the growth rate is 1 mm/week, and the judgment is performed according to the standard, when the thalamus double apical diameter is abnormal, an abnormal prompt can be given on the ultrasonic main interface, and a doctor can make treatment according to the prompt.
Step 11: and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the target cerebellar contour of the cerebellar standard section and the second brain midline.
In one embodiment, step 11 may specifically include:
step 110: determining a slope of a cerebellar transverse diameter line based on the second cerebral midline.
It will be appreciated that the second midline brain and the cerebellar transverse diameter line are in a perpendicular relationship, and the slope of the double apical diameter line can be determined based on the perpendicular relationship.
Step 111: and determining the cerebellar transverse diameter straight line based on the slope of the cerebellar transverse diameter straight line.
In one embodiment, the apex and nadir of the cerebellar contour may be targeted, and the midpoint of the apex and nadir may be determined, from which the cerebellar transverse diameter line is determined according to its slope.
Step 112: and determining a cerebellar transverse diameter line segment based on the cerebellar transverse diameter straight line and the target cerebellar contour.
In one embodiment, the intersection point of the cerebellar transverse diameter straight line and the target cerebellar contour may be determined, and the intersection point may be connected to obtain a cerebellar transverse diameter line segment. In another embodiment, the intersection of the cerebellar transverse diameter line and the mask of the target cerebellar contour may be determined, resulting in a cerebellar transverse diameter line segment.
Step 113: and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the cerebellar transverse diameter line segment.
In one embodiment, step 113 may specifically include:
step 1130: and determining a second area range and a third area range respectively based on two end points of the cerebellar transverse diameter line segment. The top end corresponds to the second area range and the bottom end corresponds to the third area range.
Step 1131: and respectively intercepting a second region image and a third region image from the cerebellum standard tangent plane based on the second region range and the third region range.
Step 1132: and respectively determining cerebellum local contour in the second area image and the third area image.
For example, two square regions are determined as the second region range and the third region range with the two endpoints of the cerebellar transverse diameter line segment as the center and the second preset size. And taking out the square area image of the corresponding position of the cerebellum standard section, carrying out threshold segmentation, and taking the maximum contour as the local contour of the cerebellum.
Step 1133: and determining the intersection point of the cerebellum transverse diameter straight line and the cerebellum local contour.
Step 1134: and determining a final cerebellar transverse diameter line segment based on the intersection point.
It can be understood that there are two intersections between the cerebellar transverse diameter line and each cerebellar local contour, the top end point of the final cerebellar transverse diameter line segment is the upper one of the two intersections between the cerebellar transverse diameter line and the cerebellar local contour in the second region image, and the bottom end point of the final cerebellar transverse diameter line segment is the lower one of the two intersections between the cerebellar transverse diameter line and the cerebellar local contour in the third region image. The final cerebellar transverse diameter line segment is then drawn on the image, as shown in fig. 3 (b). Thus, the problem that the final cerebellum transverse diameter is short due to the fact that the result of the segmentation model is small can be avoided.
Step 1135: and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the final cerebellar transverse diameter line segment.
The method and the device can determine the length of the final cerebellum transverse diameter line segment as the cerebellum transverse diameter measurement result of the cerebellum standard section. And further, displaying the lateral diameter measurement result of the cerebellum on an ultrasonic equipment interface. It should be noted that before 24 weeks of pregnancy, the cerebellum diameter value is approximately equal to the number of pregnancy weeks, the growth rate is 1-2 mm/week for 20-38 weeks of pregnancy, and the growth rate is 0.7 mm/week for 38 weeks of pregnancy, and the judgment is performed according to the standard, when the cerebellum diameter is abnormal, an abnormal prompt can be given on the ultrasonic main interface, and a doctor can perform treatment according to the prompt.
Therefore, the cerebral midline information is combined in the process of calculating the double apical diameter and the cerebellum transverse diameter, the correct directions of the double apical diameter and the cerebellum transverse diameter are ensured, no deviation occurs, and the measurement precision is improved.
Further, the present application may further determine the lateral ventricle posterior angle inner diameter measurement result of the lateral brain standard cut plane based on the target lateral ventricle contour of the lateral brain standard cut plane. The method specifically comprises the following steps:
step 21: determining location information of a choroid plexus in a lateral ventricle based on the target lateral ventricle contour.
Further, step 21 may specifically include:
step 210: determining an outgoing lateral ventricle image from the lateral brain standard slice based on the target lateral ventricle contour.
In one embodiment, a minimum bounding rectangle of the target lateral ventricle contour may be determined, based on which the outgoing lateral ventricle image is determined from the standard lateral brain slice.
Step 211: and multiplying the mask of the target lateral ventricle outline and the local lateral ventricle image to obtain a target lateral ventricle image.
It will be appreciated that this results in a lateral ventricle image that is a black region except within the lateral ventricle contour.
Step 212: and thinning the mask of the target lateral ventricle outline to obtain a first central axis.
Step 213: and multiplying the first central axis by the target lateral ventricle image to obtain a second central axis of the choroid plexus.
Step 214: and respectively carrying out straight line fitting on the first central axis and the second central axis to obtain a first line segment and a second line segment.
Step 215: determining location information of the choroid plexus in a lateral ventricle based on the first and second line segments.
In one embodiment, a first distance between two left endpoints of the first line segment and the second line segment and a second distance between two right endpoints may be calculated, respectively, and if the first distance is less than the second distance, a left position of the choroid plexus in the lateral ventricle is determined, and if the first distance is greater than the second distance, a right position of the choroid plexus in the lateral ventricle is determined.
Step 22: determining a lateral ventricles posterior horn inner diameter line based on the position information.
In one embodiment, if the choroid plexus is located at a left position in the lateral ventricle, the right end point of the second line segment defines a perpendicular to the second line segment as the lateral ventriculo-posterior angle inner diameter straight line. And if the choroid plexus is in the right position in the lateral ventricle, determining the perpendicular line of the second line segment as the inner diameter straight line of the posterior horn of the lateral ventricle by the left endpoint of the second line segment.
Step 23: determining a lateral ventriculo-caudal posterior diameter line segment based on the lateral ventriculo-medial diameter line and the target lateral ventriculo-ventricular profile.
In one embodiment, the intersection of the inner lateral ventricle diameter line and the target lateral ventricle contour may be determined and connected to obtain a lateral ventricle posterior angle inner diameter line segment. In another embodiment, the intersection of the inner lateral ventricle diameter line and the mask of the target lateral ventricle contour may be determined, resulting in a lateral ventricle posterior horn diameter line segment. The lateral ventricles posterior horn inner diameter line segment is then plotted on the image, as shown in fig. 4 (b).
Step 24: and determining the inner diameter measurement result of the lateral ventricle posterior angle of the lateral brain standard tangent plane based on the inner diameter line segment of the lateral ventricle posterior angle.
The length of the inner diameter line segment of the lateral ventricle posterior horn can be determined as the inner diameter measurement result of the lateral ventricle posterior horn of the lateral brain standard tangent plane in the embodiment of the application. Further, the lateral ventricles posterior horn inner diameter measurement is displayed on the ultrasound device interface. It is noted that when the inner diameter of the lateral ventricle is within the range of 10-15mm, a prompt of 'slight ventricular dilatation' can be given on the ultrasonic main interface, and when the inner diameter of the lateral ventricle is larger than 15mm, a prompt of 'severe ventricular dilatation or hydrocephalus' can be given on the ultrasonic main interface, and a doctor can make treatment according to the prompt.
For example, referring to fig. 5, fig. 5 is a schematic view illustrating a specific ultrasound image processing disclosed in the embodiment of the present application. The method comprises the steps that an ultrasonic doctor clicks to start scanning, an ultrasonic probe scans the abdomen of a pregnant woman to obtain an ultrasonic video, an ultrasonic image in the ultrasonic video is input into a standard section identification model, a fetal head standard section in the ultrasonic video is identified and automatically stored, the fetal head standard section comprises a thalamus standard section, a cerebellum standard section and a lateral brain standard section, then the fetal head standard section is input into a segmentation model, a target thalamus contour and a brain midline of the thalamus standard section, a target cerebellum contour brain midline of the cerebellum standard section and a target lateral ventricle contour of the lateral brain standard section are obtained and are drawn on an image, then a double vertex diameter line segment, a cerebellum transverse diameter line segment and a lateral ventricle rear angle inner diameter line segment are determined and drawn on the image, and a measurement result is obtained.
Further, in a specific embodiment, the standard section of the fetal head is input into the segmentation model to obtain the inference results of each of the three sections, and the inference results are respectively post-processed.
The method comprises the steps of performing linear fitting on a brain midline reasoning result to obtain a brain midline, obtaining a slope of a double-vertex-diameter straight line according to a vertical relation, calculating a mass center of a thalamus contour, taking a region of 10 pixels on the left and right by taking the mass center as a center, traversing and drawing the double-vertex-diameter straight line in the region, taking an intersection of the double-vertex-diameter straight line and a mask of the thalamus contour to obtain a plurality of line segments, calculating the length of the line segments, and taking the longest line segment as the double-vertex-diameter line segment. According to the clinical requirement, the bottom end of the double-top-diameter line segment cannot penetrate through the inferior bone ring, so that a square area with a first preset size is taken at the bottom end point of the double-top-diameter line segment, a square area image of the corresponding position of the thalamus standard tangent plane is taken, threshold segmentation is carried out on the image, the maximum outline is taken to obtain the inferior bone ring outline, and the lowest point of the inferior bone ring outline is calculated to be used as the final bottom end point of the double-top-diameter line segment. If the lower bone ring contour cannot be found, the bottom end point of the double-top-diameter line segment is moved up by 10 pixels to obtain the final double-top-diameter line segment.
And for the cerebellum standard section, performing straight line fitting on the reasoning result of the cerebellum central line to obtain the cerebellum central line, then obtaining the slope of the cerebellum transverse diameter straight line according to the vertical relation, then calculating the highest point and the lowest point of the cerebellum contour, calculating the middle point of the highest point and the lowest point, drawing the cerebellum transverse diameter straight line through the middle point, and taking the intersection line segment of the cerebellum transverse diameter straight line and the mask of the cerebellum contour as the cerebellum transverse diameter line segment. In order to avoid that the final cerebellum transverse diameter line segment is short due to the fact that the result of network segmentation is small, a square area is taken from each of two end points of the cerebellum transverse diameter line segment, a square area image at the corresponding position of a cerebellum standard section is taken out, threshold segmentation is carried out on the square area image, the maximum outline is taken to obtain a cerebellum local outline, the intersection point of a cerebellum transverse diameter straight line and the cerebellum local outline is calculated, and the final cerebellum transverse diameter line segment is determined.
For the standard lateral ventricle section, the corresponding lateral ventricle local original image is obtained according to the lateral ventricle contour deduced by the network, multiplying the lateral ventricle contour map by the mask of the lateral ventricle contour map to obtain a lateral ventricle contour map of a black area around the lateral ventricle contour map, thinning the mask of the lateral ventricle contour map to obtain a first central axis and performing straight line fitting to obtain a first line segment, multiplying the lateral ventricle contour map by the thinned first central axis to obtain a central axis of a choroid plexus, performing straight line fitting to obtain a second line segment, and calculating the distance between the left end point of the first line segment and the left end point of the second line segment, and the distance between the right end points, if the distance between the left end points is small, and then, the choroid plexus is on the left side of the lateral ventricle, and a perpendicular line, namely a lateral ventricle inner diameter straight line, is drawn at the right end point of the second line segment, otherwise, a lateral ventricle posterior angle inner diameter straight line is drawn in the same way, and the intersection of the lateral ventricle posterior angle inner diameter straight line and the mask of the lateral ventricle contour is the final lateral ventricle posterior angle inner diameter line segment.
That is, in the process of examining the fetus by the sonographer, the embodiment of the present application judges in real time by a deep learning manner according to the key structure of the standard section of the thalamus, cerebellum and lateral brain of the fetus, automatically obtains and stores the standard section, outputs the contour of the thalamus, cerebellum and lateral brain, and the positions of the parietal diameter of the thalamus, the lateral diameter of the cerebellum and the internal diameter of the lateral cerebral posterior horn through a segmentation model, and automatically calculates the measurement items such as the length of the parietal diameter line, the length of the cerebellum lateral diameter line, the length of the internal diameter line of the lateral cerebral ventricular posterior horn, the perimeter area of the contour, etc., to realize automatic measurement, and displays the measurement result on the interface of the ultrasound device, and prompts the abnormal result. The cerebral midline information is combined in the process of calculating the double apical diameter and the cerebellum transverse diameter, the double apical diameter and the cerebellum transverse diameter are ensured to be correct in direction, deviation is avoided, the measurement precision is improved, a doctor can measure multiple indexes of the fetal cranium by one key, the operation is convenient, and the examination efficiency of the doctor is improved.
Therefore, the method and the device for determining the standard fetal head section in the ultrasonic video recognize the standard fetal head section in the ultrasonic video, then input the standard fetal head section into an image segmentation model, obtain the segmentation result of the standard fetal head section, wherein the segmentation result comprises the contour of the key structure in the fetal head and the brain midline, and finally determine the measurement result of the standard fetal head section based on the contour of the key structure and the brain midline. That is, according to the method, the image segmentation model is firstly utilized to determine the outline of the key structure in the head of the fetus and the brain midline, then the brain midline is utilized to assist automatic measurement, the direction accuracy of a measurement item in the standard section of the head of the fetus is guaranteed, and the accuracy of the measurement of the standard section of the head of the fetus can be improved.
Referring to fig. 6, an ultrasound image processing apparatus according to an embodiment of the present application includes:
the standard section module 11 is used for identifying a standard section of the fetal head in an ultrasonic video or identifying the standard section of the fetal head in an ultrasonic real-time mode;
the image segmentation module 12 is configured to input the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises the contour of a key structure in the fetal head and the brain midline;
and the image measuring module 13 is used for determining the measurement result of the standard fetal head section based on the key structure contour and the brain midline.
Therefore, the method and the device for determining the standard fetal head section in the ultrasonic video recognize the standard fetal head section in the ultrasonic video, then input the standard fetal head section into an image segmentation model, obtain the segmentation result of the standard fetal head section, wherein the segmentation result comprises the contour of the key structure in the fetal head and the brain midline, and finally determine the measurement result of the standard fetal head section based on the contour of the key structure and the brain midline. That is, according to the method, the image segmentation model is firstly utilized to determine the outline of the key structure in the head of the fetus and the brain midline, then the brain midline is utilized to assist automatic measurement, the direction accuracy of a measurement item in the standard section of the head of the fetus is guaranteed, and the accuracy of the measurement of the standard section of the head of the fetus can be improved.
Wherein the fetal head standard section comprises a thalamus standard section and a cerebellum standard section, the key structure contour comprises a target thalamus contour and a target cerebellum contour, and the brain midline comprises a first brain midline and a second brain midline; correspondingly, the image measuring module 13 comprises
A thalamic standard section measurement sub-module for determining a thalamic double apical diameter measurement of the thalamic standard section based on a target thalamic contour of the thalamic standard section and the first brain midline;
and the cerebellum standard section measuring submodule is used for determining a cerebellum transverse diameter measuring result of the cerebellum standard section based on the target cerebellum contour of the cerebellum standard section and the second brain midline.
In one embodiment, the thalamus standard section measurement submodule specifically includes:
a double apical diameter straight line slope determination unit for determining the slope of a double apical diameter straight line based on the first cerebral midline;
the double-top-diameter straight line determining unit is used for determining a double-top-diameter straight line based on the slope of the double-top-diameter straight line;
a double-vertex-diameter line segment determination unit for determining a double-vertex-diameter line segment based on the double-vertex-diameter straight line and the target thalamus contour;
and the thalamic double apical diameter measurement result determining unit is used for determining the thalamic double apical diameter measurement result of the thalamic standard tangent plane based on the double apical diameter line segment.
Further, the unit for determining straight lines based on double top diameters is specifically configured to determine a centroid of the target thalamus contour, and determine a target pixel range based on the centroid; the double-top-diameter straight line determining unit determines a plurality of double-top-diameter straight lines based on the target pixel range and the slope of the double-top-diameter straight lines. Correspondingly, the double-vertex-diameter line segment determining unit is specifically used for determining a plurality of line segments based on the plurality of double-vertex-diameter straight lines and the target thalamus contour; and determining the longest line segment in the line segments as a double-top-diameter line segment.
Further, the thalamus double-top diameter measurement result determining unit is specifically used for determining a lower bone ring contour from the thalamus standard tangent plane, and determining the lowest point of the lower bone ring contour as the final bottom end point of the double-top diameter line segment to obtain a final double-top diameter line segment; determining a thalamic double apical diameter measurement of the thalamic standard section based on the final double apical diameter line segment. And, said determining a contour of the inferior bone annulus from said thalamic standard cut plane comprises: determining a first area range based on the bottom end point of the double-top-diameter line segment; intercepting a first region image from the thalamic standard section based on the first region range; an inferior bone ring contour is determined from the first region image.
In one embodiment, the cerebellum standard section measurement submodule is configured to include:
a cerebellar transverse diameter straight line slope determining unit for determining the slope of the cerebellar transverse diameter straight line based on the second cerebral midline;
a cerebellum transverse diameter straight line determining unit for determining a cerebellum transverse diameter straight line based on the slope of the cerebellum transverse diameter straight line;
a cerebellum transverse diameter line segment determining unit, which is used for determining the cerebellum transverse diameter line segment based on the cerebellum transverse diameter straight line and the target cerebellum contour;
and the cerebellum transverse diameter measurement result determining unit is used for determining the cerebellum transverse diameter measurement result of the cerebellum standard section based on the cerebellum transverse diameter line segment.
Further, the cerebellar transverse diameter measurement result determining unit is specifically configured to determine a second area range and a third area range based on two end points of the cerebellar transverse diameter line segment, respectively; respectively intercepting a second area image and a third area image from the cerebellum standard tangent plane based on the second area range and the third area range; respectively determining cerebellum local contours in the second area image and the third area image; determining the intersection point of the cerebellum transverse diameter straight line and the cerebellum local contour; determining a final cerebellum transverse diameter line segment based on the intersection point; and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the final cerebellar transverse diameter line segment.
Further, the standard fetal head section further comprises a lateral brain standard section, and the key structure contour further comprises a target lateral ventricle contour; correspondingly, the image measuring module 13 further includes:
and the lateral cerebral standard section measuring submodule is used for determining the internal diameter measurement result of the lateral ventricle posterior horn of the lateral cerebral standard section based on the target lateral ventricle contour of the lateral cerebral standard section.
In one embodiment, the lateral brain standard section measurement submodule specifically includes:
a choroid plexus position determination unit for determining position information of the choroid plexus in the lateral ventricle based on the target lateral ventricle contour;
a ventricles posterior horn inner diameter straight line determining unit for determining a ventricles posterior horn inner diameter straight line based on the position information;
a ventriculo-ventriculus posterior angle inner diameter line segment determining unit, which is used for determining the ventriculo-ventriculus posterior angle inner diameter line segment based on the ventriculo-ventriculus inner diameter straight line and the target ventriculus-ventriculus outline;
and the lateral ventricle posterior angle inner diameter measurement result determining unit is used for determining the lateral ventricle posterior angle inner diameter measurement result of the lateral brain standard tangent plane based on the lateral ventricle posterior angle inner diameter line segment.
Wherein the choroid plexus position determination unit is specifically configured to determine an external lateral ventricle image from the lateral brain standard cut plane based on the target lateral ventricle contour; multiplying the mask of the target lateral ventricle outline and the local lateral ventricle image to obtain a target lateral ventricle image; performing thinning processing on the mask of the target lateral ventricle outline to obtain a first central axis; multiplying the first central axis by the target lateral ventricle image to obtain a second central axis of the choroid plexus; respectively performing straight line fitting on the first central axis and the second central axis to obtain a first line segment and a second line segment; determining location information of the choroid plexus in a lateral ventricle based on the first and second line segments.
In one embodiment, the training process of the image segmentation model includes: acquiring a first training sample set; the first training sample set comprises a fetal head standard section sample and label information corresponding to the fetal head standard section sample; the fetal head standard section sample comprises a thalamus standard section sample, a cerebellum standard section sample and a lateral brain standard section sample; the label information of the thalamus standard section sample comprises thalamus contour lines and brain midline, the label information of the cerebellum standard section sample comprises cerebellum contour lines and brain midline, and the label information of the lateral cerebrum standard section comprises lateral ventricle contour lines; inputting the fetal head standard section sample into a first initial model to obtain a thalamus contour, a cerebellum contour, a brain midline and a lateral ventricle contour corresponding to the fetal head standard section sample; calculating a training loss based on the thalamic contour, cerebellar contour, the midline of the brain, the lateral ventricle contour, and label information of the fetal head standard sectional sample, and performing parameter adjustment on the first initial model based on the training loss; and if the condition that the training completion condition is met is detected, determining the first initial model after parameter adjustment as an image segmentation model.
In one embodiment, the standard section module 11 is specifically configured to identify a standard section of the fetal head in an ultrasound video, or identify a standard section of the fetal head in an ultrasound real-time mode, using a standard section identification model;
the standard section recognition model is obtained by training a second initial model by using a second training sample set, wherein the second training sample set comprises a fetal head standard section sample, a non-fetal head standard section sample and label information.
Referring to fig. 7, the embodiment of the present application discloses an ultrasound apparatus 20, which includes a processor 21 and a memory 22; wherein, the memory 22 is used for saving computer programs; the processor 21 is configured to execute the computer program and the ultrasound image processing method disclosed in the foregoing embodiments.
For the specific process of the ultrasound image processing method, reference may be made to the corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
The memory 22 is used as a carrier for resource storage, and may be a read-only memory, a random access memory, a magnetic disk or an optical disk, and the storage mode may be a transient storage mode or a permanent storage mode.
In addition, the ultrasound apparatus 20 further includes a power supply 23, a communication interface 24, an input-output interface 25, and a communication bus 26; the power supply 23 is configured to provide an operating voltage for each hardware device on the ultrasound device 20; the communication interface 24 can create a data transmission channel between the ultrasound apparatus 20 and an external device, and a communication protocol followed by the communication interface is any communication protocol applicable to the technical solution of the present application, and is not specifically limited herein; the input/output interface 25 is configured to obtain external input data or output data to the outside, and a specific interface type thereof may be selected according to a specific application requirement, which is not specifically limited herein.
Of course, in another embodiment, an electronic device may be provided that includes a processor and a memory; wherein the memory is used for storing a computer program; the processor is configured to execute the computer program and the ultrasound image processing method disclosed in the foregoing embodiment. The electronic equipment is connected with the ultrasonic equipment, obtains the ultrasonic video through communication with the ultrasonic equipment and carries out corresponding processing.
That is, the computer program implemented by the method of the present application can be embedded into the data stream of the existing ultrasound imaging system, and the ultrasound video to be analyzed and processed is directly called from the data stream of the existing ultrasound system; the method can also be used as an independent system program and placed in a computer with storage and operation functions, and the ultrasonic video to be processed is obtained through communication with the ultrasonic imaging system or connection with a data cloud, so that independent reasoning and operation are carried out.
Further, the present application also discloses a computer readable storage medium for storing a computer program, wherein the computer program is executed by a processor to implement the ultrasound image processing method disclosed in the foregoing embodiments.
For the specific process of the ultrasound image processing method, reference may be made to the corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
In the present specification, the embodiments are described in a progressive manner, and each embodiment focuses on differences from other embodiments, and the same or similar parts between the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The above method, apparatus, device and medium for processing an ultrasound image provided by the present application are described in detail, and a specific example is applied in the description to explain the principle and the implementation of the present application, and the description of the above embodiment is only used to help understand the method and the core idea of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.
Claims (16)
1. An ultrasound image processing method, comprising:
identifying a standard fetal head section in an ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode;
inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises a contour of a key structure in the fetal head and a brain midline;
determining a measurement of the fetal head standard section based on the critical structure contour and the brain midline.
2. The method of claim 1, wherein the standard section of the fetal head includes a standard section of thalamus and a standard section of cerebellum, the contour of the critical structure includes a target contour of thalamus and a target contour of cerebellum, and the brain midline includes a first brain midline and a second brain midline; correspondingly, the determining the measurement result of the standard section of the fetal head based on the contour of the key structure and the brain midline comprises:
determining a thalamic double apical diameter measurement of the thalamic standard section based on a target thalamic contour of the thalamic standard section and the first brain centerline;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the target cerebellar contour of the cerebellar standard section and the second brain midline.
3. The method of claim 2, wherein said determining a thalamic double apical diameter measurement of said thalamic standard cut based on said first brain centerline and a target thalamic contour of said thalamic standard cut comprises:
determining a slope of a dual apical diameter line based on the first cerebral midline;
determining a double-top-diameter straight line based on the slope of the double-top-diameter straight line;
determining a double apical diameter line segment based on the double apical diameter line and the target thalamic contour;
determining thalamic double apical diameter measurements of the thalamic standard section based on the double apical diameter line segments.
4. The method according to claim 3, wherein the determining a double-vertex-diameter line based on a slope of the double-vertex-diameter line; determining a double apical diameter line segment based on the double apical diameter line and the target thalamic contour, including:
determining a centroid of the target thalamic contour and determining a target pixel range based on the centroid;
determining a plurality of double-top-diameter straight lines based on the target pixel range and the slope of the double-top-diameter straight lines;
determining a plurality of line segments based on the plurality of double apical diameter lines and the target thalamic contour;
and determining the longest line segment in the line segments as a double-top-diameter line segment.
5. The method of claim 3, wherein said determining a thalamic double apical diameter measurement of said thalamic standard tangent plane based on said double apical diameter line segment comprises:
determining a lower bone ring contour from the thalamus standard tangent plane, and determining the lowest point of the lower bone ring contour as the final bottom end point of the double-top diameter line segment to obtain a final double-top diameter line segment;
determining a thalamic double apical diameter measurement of the thalamic standard section based on the final double apical diameter line segment.
6. The method of claim 5, wherein said determining the contour of the inferior bone ring from the standard thalamic tangent plane comprises:
determining a first area range based on the bottom end point of the double-top-diameter line segment;
intercepting a first region image from the thalamic standard section based on the first region range;
an inferior bone ring contour is determined from the first region image.
7. The method of claim 2, wherein determining the cerebellar transverse diameter measurement of the cerebellar standard cut based on the target cerebellar contour of the cerebellar standard cut and the second brain centerline comprises:
determining a slope of a cerebellar transverse diameter line based on the second cerebral midline;
determining a cerebellar transverse diameter line based on the slope of the cerebellar transverse diameter line;
determining a cerebellar transverse diameter line segment based on the cerebellar transverse diameter straight line and the target cerebellar contour;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the cerebellar transverse diameter line segment.
8. The method of claim 7, wherein said determining a cerebellar transverse diameter measurement of the cerebellar standard section based on the cerebellar transverse diameter line segment comprises:
determining a second area range and a third area range respectively based on two end points of the cerebellar transverse diameter line segment;
respectively intercepting a second area image and a third area image from the cerebellum standard tangent plane based on the second area range and the third area range;
respectively determining cerebellum local contours in the second area image and the third area image;
determining the intersection point of the cerebellum transverse diameter straight line and the cerebellum local contour;
determining a final cerebellar transverse diameter line segment based on the intersection points;
and determining a cerebellar transverse diameter measurement result of the cerebellar standard section based on the final cerebellar transverse diameter line segment.
9. The method of claim 1, wherein the fetal head standard section further comprises a lateral brain standard section, and the critical structure contour further comprises a target lateral ventricle contour; correspondingly, the method further comprises the following steps:
and determining the inner diameter measurement result of the lateral ventricle posterior horn of the lateral brain standard section based on the target lateral ventricle contour of the lateral brain standard section.
10. The method of claim 9, wherein the determining a lateral ventriculo-posterior angle inner diameter measurement of the lateral brain standard cut plane based on the target lateral ventricle contour of the lateral brain standard cut plane comprises:
determining location information of a choroid plexus in a lateral ventricle based on the target lateral ventricle contour;
determining a lateral ventricle posterior angle inner diameter straight line based on the position information;
determining a lateral ventriculo-caudal posterior radius line segment based on the lateral ventriculo-medial diameter line and the target lateral ventriculo-ventricular profile;
and determining the lateral ventricle posterior angle inner diameter measurement result of the lateral ventricle posterior angle standard tangent plane based on the lateral ventricle posterior angle inner diameter line segment.
11. The method of ultrasound image processing according to claim 10, wherein said determining location information of choroid plexus in lateral ventricle based on said target lateral ventricle contour comprises:
determining an outgoing lateral ventricle image from the lateral brain standard slice based on the target lateral ventricle contour;
multiplying the mask of the target lateral ventricle outline and the local lateral ventricle image to obtain a target lateral ventricle image;
performing thinning processing on the mask of the target lateral ventricle outline to obtain a first central axis;
multiplying the first central axis by the target lateral ventricle image to obtain a second central axis of the choroid plexus;
respectively performing straight line fitting on the first central axis and the second central axis to obtain a first line segment and a second line segment;
determining location information of the choroid plexus in a lateral ventricle based on the first and second line segments.
12. The method of claim 1, wherein the training process of the image segmentation model comprises:
acquiring a first training sample set; the first training sample set comprises a fetal head standard section sample and label information corresponding to the fetal head standard section sample; the fetal head standard section sample comprises a thalamus standard section sample, a cerebellum standard section sample and a lateral brain standard section sample; the label information of the thalamus standard section sample comprises thalamus contour lines and brain midline, the label information of the cerebellum standard section sample comprises cerebellum contour lines and brain midline, and the label information of the lateral cerebrum standard section comprises lateral ventricle contour lines;
inputting the fetal head standard section sample into a first initial model to obtain a thalamus contour, a cerebellum contour, a brain midline and a lateral ventricle contour corresponding to the fetal head standard section sample;
calculating a training loss based on the thalamic contour, cerebellar contour, the midline of the brain, the lateral ventricle contour, and label information of the fetal head standard sectional sample, and performing parameter adjustment on the first initial model based on the training loss;
and if the condition that the training completion condition is met is detected, determining the first initial model after parameter adjustment as an image segmentation model.
13. The method of any one of claims 1 to 12, wherein the identifying a standard fetal head slice in the ultrasound video or in the ultrasound real-time mode comprises:
identifying a standard fetal head section in an ultrasonic video or identifying the standard fetal head section in an ultrasonic real-time mode by using a standard section identification model;
the standard section recognition model is obtained by training a second initial model by using a second training sample set, wherein the second training sample set comprises a fetal head standard section sample, a non-fetal head standard section sample and label information.
14. An ultrasound image processing apparatus characterized by comprising:
the standard section module is used for identifying a standard section of the head of the fetus in an ultrasonic video or identifying the standard section of the head of the fetus in an ultrasonic real-time mode;
the image segmentation module is used for inputting the standard fetal head section into an image segmentation model to obtain a segmentation result of the standard fetal head section; wherein the segmentation result comprises a contour of a key structure in the fetal head and a brain midline;
and the image measuring module is used for determining the measurement result of the standard fetal head section based on the key structure contour and the brain midline.
15. An ultrasound device comprising a processor and a memory; wherein the content of the first and second substances,
the memory is used for storing a computer program;
the processor for executing the computer program to implement the ultrasound image processing method of any one of claims 1 to 13.
16. A computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the ultrasound image processing method of any of claims 1 to 13.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2022107600612 | 2022-06-30 | ||
CN202210760061 | 2022-06-30 |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115063395A true CN115063395A (en) | 2022-09-16 |
Family
ID=83203714
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210801337.7A Pending CN115063395A (en) | 2022-06-30 | 2022-07-08 | Ultrasonic image processing method, device, equipment and medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115063395A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115919367A (en) * | 2022-12-09 | 2023-04-07 | 开立生物医疗科技(武汉)有限公司 | Ultrasonic image processing method and device, electronic equipment and storage medium |
-
2022
- 2022-07-08 CN CN202210801337.7A patent/CN115063395A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115919367A (en) * | 2022-12-09 | 2023-04-07 | 开立生物医疗科技(武汉)有限公司 | Ultrasonic image processing method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11229419B2 (en) | Method for processing 3D image data and 3D ultrasonic imaging method and system | |
CN108520519B (en) | Image processing method and device and computer readable storage medium | |
CN102458222B (en) | Image processing apparatus and image processing method | |
US20050063576A1 (en) | System and method for utilizing shape analysis to assess fetal abnormality | |
CN111368586B (en) | Ultrasonic imaging method and system | |
CN111374712B (en) | Ultrasonic imaging method and ultrasonic imaging equipment | |
CN109727240B (en) | Method and related device for stripping shielding tissues of three-dimensional ultrasonic image | |
CN113077479A (en) | Automatic segmentation method, system, terminal and medium for acute ischemic stroke focus | |
CN110279433A (en) | A kind of fetus head circumference automatic and accurate measurement method based on convolutional neural networks | |
CN111899253B (en) | Abnormal judgment analysis method and device for fetal craniocerebral section image | |
WO2022062459A1 (en) | Automatic measurement method and device for fetal structural characteristic | |
CN112349391A (en) | Optimized rib automatic labeling method | |
US20240225587A1 (en) | Ultrasound automated detection and display method of cranial abnormal regions | |
CN115063395A (en) | Ultrasonic image processing method, device, equipment and medium | |
CN115690556B (en) | Image recognition method and system based on multi-mode imaging features | |
CN113409275B (en) | Method for determining thickness of transparent layer behind fetal neck based on ultrasonic image and related device | |
US20220249060A1 (en) | Method for processing 3d image data and 3d ultrasonic imaging method and system | |
CN115861347A (en) | Method and device for extracting focus region, electronic equipment and readable storage medium | |
CN112767314A (en) | Medical image processing method, device, equipment and storage medium | |
CN118039088B (en) | Artificial intelligence auxiliary diagnosis data processing system | |
Nabila et al. | Automated Cerebral Lateral Ventricle Ratio Measurement From 2-Dimensional Fetal Ultrasound Image to Predict Ventriculomegaly | |
CN114880961B (en) | Hydrodynamics-based vessel segmentation and naming method | |
CN112638267A (en) | Ultrasonic imaging method and system, storage medium, processor and computer device | |
WO2022141085A1 (en) | Ultrasonic detection method and ultrasonic imaging system | |
CN114334095A (en) | Intelligent identification method and system for ultrasonic examination and terminal equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |