CN116091851A - Quick classification method for ultrasonic cardiac apex tangent plane image - Google Patents
Quick classification method for ultrasonic cardiac apex tangent plane image Download PDFInfo
- Publication number
- CN116091851A CN116091851A CN202310386731.3A CN202310386731A CN116091851A CN 116091851 A CN116091851 A CN 116091851A CN 202310386731 A CN202310386731 A CN 202310386731A CN 116091851 A CN116091851 A CN 116091851A
- Authority
- CN
- China
- Prior art keywords
- image
- apex
- cavity
- heart
- template image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 22
- 230000000747 cardiac effect Effects 0.000 title claims description 30
- 239000013598 vector Substances 0.000 claims abstract description 57
- 238000012545 processing Methods 0.000 claims abstract description 31
- 238000002604 ultrasonography Methods 0.000 claims description 24
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012549 training Methods 0.000 abstract description 6
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002107 myocardial effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention relates to the technical field of ultrasonic image processing, in particular to a rapid classification method of an ultrasonic heart apex tangent plane image. The technical scheme is as follows: and carrying out feature enhancement on the ultrasonic heart apex tangent plane image, acquiring a feature vector of the ultrasonic heart apex tangent plane image in a coding and decoding mode, and judging the type of the ultrasonic heart apex tangent plane image by comparing the proximity degree of the feature vector of the ultrasonic heart apex tangent plane image and the feature vector of the heart apex two-cavity tangent plane template image, the heart apex three-cavity tangent plane template image and the heart apex four-cavity tangent plane template image. The invention has the beneficial effects that: the method can realize the rapid classification of the ultrasonic heart apex tangent plane image, and has the advantages of no need of a large number of samples for model training and less memory occupation in the processing process.
Description
Technical Field
The invention relates to the technical field of ultrasonic image processing, in particular to a rapid classification method of an ultrasonic heart apex tangent plane image.
Background
The cardiac ejection fraction (Ejection Fraction, EF) and the myocardial Strain Rate (SR) in cardiac ultrasound examination are important indicators for assessing cardiac functional states, and a doctor is required to hold a probe to acquire ultrasound images of a four-chamber section of the apex, a three-chamber section of the apex and a two-chamber section of the apex. And (5) counting different section results to finally obtain corresponding indexes. Thus requiring the physician to manually select which of the cardiac cusps currently belongs to. The automatic identification of the apex section can reduce the operation of a doctor on the instrument, assist the doctor in checking, and improve the working efficiency. Most of the prior classification technologies use deep learning training models to realize classification, the model training needs a large number of samples, the model training takes a long time, the requirements on the processing performance of the system are high, the system needs to occupy a large memory, and the processing time is long.
Disclosure of Invention
The invention aims to provide a method for rapidly classifying an ultrasonic heart apex section image, in particular to a method which does not need a training model, occupies little memory and can rapidly classify the ultrasonic heart apex section image.
In order to achieve the above purpose, the invention adopts the following technical scheme: a rapid classification method of ultrasonic heart apex section images comprises the following steps:
s01, inputting an ultrasonic heart apex section image as an input image, and adjusting the input image to a preset length and width value and cutting the input image to a preprocessing image of the input image.
S02, performing image compression on the preprocessed image obtained in the step S01, performing feature enhancement processing on the compressed image to enhance the contour features of the image, reducing the non-edge features of the image, and obtaining the image after the feature enhancement processing.
S03, carrying out coding processing on the image subjected to the characteristic enhancement processing in the step S02 to obtain characteristic points in the image, and then carrying out decoding conversion according to the characteristic points to obtain the characteristic vector of the ultrasonic heart apex surface image input in the step S01.
S04, judging which of the three types of the heart apex section template image is closest to the feature vector of the heart apex section template image, namely the heart apex two-cavity section template image, the heart apex three-cavity section template image and the heart apex four-cavity section template image, wherein the type of the closest heart apex section template image is the heart apex section type of the input ultrasonic heart apex section image.
Specifically, in the step S04, when the feature vector of the input image is closest to the feature vector of the heart tip section template image of the three types of heart tip section template images, i.e., the heart tip two-cavity section template image, the heart tip three-cavity section template image, and the heart tip four-cavity section template image, the distance between the feature vector of the heart tip section template image of the three types of heart tip two-cavity section template image, the heart tip three-cavity section template image, and the heart tip four-cavity section template image and the feature vector of the input image is calculated, and the type of the heart tip section template image with the minimum feature vector distance from the feature vector of the input image is the type of the heart tip section closest to the input image.
Specifically, in step S04, feature vectors of the two-cavity apex tangent plane template image, the three-cavity apex tangent plane template image, and the four-cavity apex tangent plane template image are obtained by using the two-cavity apex tangent plane template image, the three-cavity apex tangent plane template image, and the four-cavity apex tangent plane template image as input images and adopting the methods described in steps S01-S03.
Preferably, in step S02, the feature enhancement processing is performed on the compressed image using gaussian smoothing processing and hessian matrix processing.
In step S03, the image after the feature enhancement processing is encoded, and when the feature points in the image are obtained, the image features are obtained by starting to rotate counterclockwise from the center position of the image after the feature enhancement processing, and the image data is converted into the feature points according to whether the gray level value exists at the current position as the basis for judging whether the features exist.
In step S03, when the feature vector of the ultrasound cardiac apex surface image input in step S01 is obtained by performing decoding conversion according to the feature points, the feature points are converted into feature vectors by performing data type conversion on the feature points.
The invention has the beneficial effects that: the ultrasonic heart apex tangent plane image is subjected to feature enhancement, the feature vector of the ultrasonic heart apex tangent plane image is obtained in a coding and decoding mode, the type of the ultrasonic heart apex tangent plane image is judged by judging the proximity degree of the feature vector of the ultrasonic heart apex tangent plane image and the feature vector of the heart apex two-cavity tangent plane template image, the heart apex three-cavity tangent plane template image and the heart apex four-cavity tangent plane template image, the ultrasonic heart apex tangent plane image is rapidly classified, and the ultrasonic heart apex tangent plane image processing method has the advantages that a large number of samples are not required to be subjected to model training, and the memory occupied in the processing process is small.
Drawings
FIG. 1 is a flow chart of a method for rapid classification of ultrasound cardiac apex images in an embodiment.
Fig. 2 is a graph of distance between feature vectors of two-cavity section images of the apex of the ultrasound heart and feature vectors of two-cavity section template images of the apex of the ultrasound heart, three-cavity section template images of the apex of the ultrasound heart and four-cavity section template images of the apex of the ultrasound heart, which are selected in the embodiment.
FIG. 3 is a graph showing the distance between the feature vector of the selected 140 frames of the three-cavity sectional ultrasound apex image and the feature vector of the two-cavity sectional ultrasound apex template image, the three-cavity sectional ultrasound apex template image and the four-cavity sectional ultrasound apex template image.
Fig. 4 is a graph showing the distance between the feature vector of the selected 131-frame ultrasonic apex four-cavity tangent plane image and the feature vector of the ultrasonic apex two-cavity tangent plane template image, the ultrasonic apex three-cavity tangent plane template image and the ultrasonic apex four-cavity tangent plane template image.
Description of the embodiments
Embodiment 1, referring to fig. 1, a method for rapid classification of ultrasound apical section images, comprises the steps of:
s01, inputting an ultrasonic heart apex section image as an input image, and adjusting the input image to a preset length and width value and cutting the input image to a preprocessing image of the input image.
S02, performing image compression on the preprocessed image obtained in the step S01, performing feature enhancement processing on the compressed image to enhance the contour features of the image, reducing the non-edge features of the image, and obtaining the image after the feature enhancement processing. Among them, the feature enhancement processing for the compressed image preferably employs gaussian smoothing processing and hessian matrix processing.
S03, carrying out coding processing on the image subjected to the characteristic enhancement processing in the step S02 to obtain characteristic points in the image, and then carrying out decoding conversion according to the characteristic points to obtain the characteristic vector of the ultrasonic heart apex surface image input in the step S01.
S04, judging which of the three types of the heart apex section template image is closest to the feature vector of the heart apex section template image, namely the heart apex two-cavity section template image, the heart apex three-cavity section template image and the heart apex four-cavity section template image, wherein the type of the closest heart apex section template image is the heart apex section type of the input ultrasonic heart apex section image.
Specifically, in step S04, when the feature vector of the input image is determined to be closest to the feature vector of which of the three apical section template images, i.e., the apical two-cavity template image, the apical three-cavity template image, and the apical four-cavity template image, the distance between the feature vector of the three kinds of apical section template images, i.e., the apical two-cavity template image, the apical three-cavity template image, and the apical four-cavity template image, and the feature vector of the input image is calculated, respectively, and the type of the apical section template image having the smallest distance between the feature vector and the feature vector of the input image is the type of the apical section closest to the input image.
In addition, in step S04, feature vectors of the two-cavity, three-cavity and four-cavity template images are obtained by using the two-cavity, three-cavity and four-cavity template images as input images and adopting the methods described in steps S01 to S03.
In step S03, the image after the feature enhancement processing is encoded, and when the feature points in the image are obtained, the image features are obtained by starting to rotate counterclockwise from the center position of the image after the feature enhancement processing, and the image data is converted into the feature points according to whether the gray level value exists at the current position as the basis for judging whether the features exist.
In step S03, when the feature vector of the ultrasound cardiac apex surface image input in step S01 is obtained by performing decoding conversion according to the feature points, the feature points are converted into feature vectors by performing data type conversion on the feature points.
In addition, in this embodiment, 200 frames of ultrasonic apex two-cavity tangent plane images, 140 frames of ultrasonic apex three-cavity tangent plane images and 131 frames of ultrasonic apex four-cavity tangent plane images are selected for testing, fig. 2-4 are test results, and according to fig. 2, the feature vector of the ultrasonic apex two-cavity tangent plane images is closest to the feature vector of the apex two-cavity tangent plane template image, that is, the distance between the feature vector of the ultrasonic apex two-cavity apex tangent plane image and the feature vector of the apex two-cavity tangent plane template image is the smallest (corresponding to the red curve in the graph), so that the accuracy rate reaches 99%; correspondingly, according to fig. 3, the feature vector of the ultrasonic heart apex three-cavity tangent plane image is closest to the feature vector of the heart apex three-cavity tangent plane template image, namely, the distance between the feature vector of the ultrasonic heart apex three-cavity tangent plane image and the feature vector of the heart apex three-cavity tangent plane template image is minimum (corresponding to a green curve in fig. 3), and the accuracy rate reaches 99%; according to fig. 4, the feature vector of the ultrasonic heart apex four-cavity tangent plane image is closest to the feature vector of the heart apex four-cavity tangent plane template image, namely, the distance between the feature vector of the ultrasonic heart apex tangent plane image of the heart apex four-cavity tangent plane and the feature vector of the heart apex four-cavity tangent plane template image is minimum (corresponding to a blue curve in the figure), and the accuracy rate reaches 97%.
Of course, the above embodiments are only preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, so that all equivalent modifications made in the principles of the present invention are included in the scope of the present invention.
Claims (6)
1. A rapid classification method of an ultrasonic heart apex section image is characterized by comprising the following steps of: the method comprises the following steps:
s01, inputting an ultrasonic heart apex section image as an input image, and adjusting the input image to a preset length and width value and cutting the input image to a preprocessing image of the input image;
s02, carrying out image compression on the preprocessed image obtained in the step S01, carrying out feature enhancement processing on the compressed image to enhance the contour features of the image, reducing the non-edge features of the image, and obtaining the image after the feature enhancement processing;
s03, carrying out coding treatment on the image subjected to the characteristic enhancement treatment in the step S02 to obtain characteristic points in the image, and then carrying out decoding conversion according to the characteristic points to obtain characteristic vectors of the ultrasonic heart apex surface image input in the step S01;
s04, judging which of the three types of the heart apex section template image is closest to the feature vector of the heart apex section template image, namely the heart apex two-cavity section template image, the heart apex three-cavity section template image and the heart apex four-cavity section template image, wherein the type of the closest heart apex section template image is the heart apex section type of the input ultrasonic heart apex section image.
2. The method for rapid classification of ultrasound cardiac apex images according to claim 1, wherein: in the step S04, when the feature vector of the input ultrasound cardiac tip section image is closest to the feature vector of which cardiac tip section template image of the three types of cardiac tip section template images, i.e., cardiac tip two-cavity section template image, cardiac tip three-cavity section template image and cardiac tip four-cavity section template image, the distance between the feature vector of the three types of cardiac tip section template image, i.e., cardiac tip two-cavity section template image, cardiac tip three-cavity section template image and cardiac tip four-cavity section template image and the feature vector of the input ultrasound cardiac tip section image is calculated, and the type of cardiac tip section template image with the minimum distance between the feature vector and the feature vector of the input ultrasound cardiac tip section image is the cardiac tip section type closest to the input ultrasound cardiac tip section image.
3. The method for rapid classification of ultrasound cardiac apex images according to claim 1, wherein: in the step S04, template feature vectors of the two-cavity heart tip section template image, the three-cavity heart tip section template image and the four-cavity heart tip section template image are obtained by using the two-cavity heart tip section template image, the three-cavity heart tip section template image and the four-cavity heart tip section template image as input images and adopting the methods described in the steps S01-S03.
4. The method for rapid classification of ultrasound cardiac apex images according to claim 1, wherein: in the step S02, the feature enhancement processing is performed on the compressed image by gaussian smoothing processing and hessian matrix processing.
5. The method for rapid classification of ultrasound cardiac apex images according to claim 1, wherein: in the step S03, the image after the feature enhancement processing is encoded, and when the feature points in the image are obtained, the image features are obtained by starting to rotate counterclockwise from the center position of the image after the feature enhancement processing, and the image data is converted into the feature points according to whether the gray level value exists at the current position as the basis for judging whether the features exist.
6. The method for rapid classification of ultrasound cardiac apex images according to claim 1, wherein: in the step S03, when the feature vector of the ultrasound cardiac apex surface image input in the step S01 is obtained by performing decoding conversion according to the feature points, the feature points are converted into feature vectors by performing data type conversion on the feature points.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310386731.3A CN116091851A (en) | 2023-04-12 | 2023-04-12 | Quick classification method for ultrasonic cardiac apex tangent plane image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310386731.3A CN116091851A (en) | 2023-04-12 | 2023-04-12 | Quick classification method for ultrasonic cardiac apex tangent plane image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116091851A true CN116091851A (en) | 2023-05-09 |
Family
ID=86188184
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310386731.3A Pending CN116091851A (en) | 2023-04-12 | 2023-04-12 | Quick classification method for ultrasonic cardiac apex tangent plane image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116091851A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6067369A (en) * | 1996-12-16 | 2000-05-23 | Nec Corporation | Image feature extractor and an image feature analyzer |
CN102663400A (en) * | 2012-04-16 | 2012-09-12 | 北京博研新创数码科技有限公司 | LBP (length between perpendiculars) characteristic extraction method combined with preprocessing |
CN112001373A (en) * | 2020-10-28 | 2020-11-27 | 北京妙医佳健康科技集团有限公司 | Article identification method and device and storage medium |
-
2023
- 2023-04-12 CN CN202310386731.3A patent/CN116091851A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6067369A (en) * | 1996-12-16 | 2000-05-23 | Nec Corporation | Image feature extractor and an image feature analyzer |
CN102663400A (en) * | 2012-04-16 | 2012-09-12 | 北京博研新创数码科技有限公司 | LBP (length between perpendiculars) characteristic extraction method combined with preprocessing |
CN112001373A (en) * | 2020-10-28 | 2020-11-27 | 北京妙医佳健康科技集团有限公司 | Article identification method and device and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898175B (en) | Computer-aided model construction method based on deep learning gastric cancer pathological section | |
CN108464840B (en) | Automatic detection method and system for breast lumps | |
CN110930418B (en) | Retina blood vessel segmentation method fusing W-net and conditional generation confrontation network | |
CN108109140A (en) | Low Grade Gliomas citric dehydrogenase non-destructive prediction method and system based on deep learning | |
CN111476754B (en) | Bone marrow cell image artificial intelligence auxiliary grading diagnosis system and method | |
CN111986101B (en) | Cerebrovascular map construction method | |
CN111598894B (en) | Retina blood vessel image segmentation system based on global information convolution neural network | |
CN110613483A (en) | Method and system for detecting fetal craniocerebral abnormality based on machine learning | |
CN112270676A (en) | Computer-aided judgment method for endometrial cancer muscle layer infiltration depth of MRI (magnetic resonance imaging) image | |
CN111242949B (en) | Fundus image blood vessel segmentation method based on full convolution neural network multi-scale features | |
CN112270667B (en) | TI-RADS-based integrated deep learning multi-tag identification method | |
CN109002831A (en) | A kind of breast density classification method, system and device based on convolutional neural networks | |
CN112907581A (en) | MRI (magnetic resonance imaging) multi-class spinal cord tumor segmentation method based on deep learning | |
CN116758336A (en) | Medical image intelligent analysis system based on artificial intelligence | |
CN116310329A (en) | Skin lesion image segmentation method based on lightweight multi-scale UNet | |
CN110327033B (en) | Myocardial infarction electrocardiogram screening method based on deep neural network | |
CN116109558A (en) | Method for screening fetal chromosomal abnormalities based on neck semitransparent thickness measurement | |
CN116091851A (en) | Quick classification method for ultrasonic cardiac apex tangent plane image | |
CN113269799A (en) | Cervical cell segmentation method based on deep learning | |
CN113344922A (en) | Method for automatically segmenting cerebral microhemorrhage points | |
CN117036288A (en) | Tumor subtype diagnosis method for full-slice pathological image | |
EP4167184A1 (en) | Systems and methods for plaque identification, plaque composition analysis, and plaque stability detection | |
CN114612421A (en) | Oval foramen ovale microbubble counting method based on deep learning | |
CN115359003A (en) | Two-step tunnel gray image crack identification method, system, medium and equipment | |
CN111466894B (en) | Ejection fraction calculation method and system based on deep learning |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230509 |
|
RJ01 | Rejection of invention patent application after publication |