Coronary heart disease echocardiogram pathological feature extraction method and system

Abstract

The invention provides a coronary heart disease echocardiogram pathological feature extraction method and system. Static structure features and myocardial function features of the heart of a target coronary heart disease patient are determined through an ultrasonic video stream of the heart of the target coronary heart disease patient; determining a region mapping matrix of a relationship between a coronary artery blood supply region of the target coronary heart disease patient and a myocardial injury state according to the static structural characteristics and the myocardial functional characteristics, and determining boundary sharpening development of a pathological region in the heart of the target coronary heart disease patient according to the region mapping matrix; acquiring a myocardial movement track of the target coronary heart disease patient, and determining a longitudinal strain gradient of a basal segment and a cardiac apex segment of the heart of the target coronary heart disease patient according to the myocardial movement track; and based on the boundary sharpening development and the longitudinal strain gradient of the pathological area and in combination with the clinical pathological data of the target coronary heart disease patient, generating the cardiac physical characteristics of the target coronary heart disease patient. According to the scheme, the static structure and the dynamic function of the heart can be effectively associated and fused to extract the coronary heart disease echocardiogram features.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of feature extraction, more particularly, the present application relates to a coronary heart disease echocardiogram pathological feature extraction method and system. BACKGROUND

[0002] Feature extraction refers to the process of identifying, separating and extracting features from raw data that effectively represent the essence of the data or key information. The purpose is to convert complex and redundant raw data into more concise and representative feature vectors, so as to facilitate subsequent data analysis, pattern recognition, machine learning and other tasks.

[0003] In the diagnosis of coronary heart disease, echocardiography is widely used due to its non-invasive, convenient and real-time dynamic observation of the heart. However, existing technologies mainly focus on analyzing the static structure (such as wall thickness, chamber size) or dynamic function indicators (such as ejection fraction, overall strain) of the heart in isolation, lacking effective methods to deeply correlate and integrate these features, especially the precise quantification of the correspondence between specific coronary artery blood supply areas and their dominated myocardial injury states. At the same time, the boundary identification of pathological areas (such as ischemic or infarct myocardium) is often not clear enough, and the overall motion pattern of the heart is not fully integrated for comprehensive evaluation, which leads to insufficient accuracy and specificity in positioning, degree evaluation and boundary delineation of coronary heart disease myocardial injury, limiting the potential of echocardiography in precise diagnosis and personalized treatment planning. Therefore, how to effectively correlate and integrate the static structure and dynamic function of the heart to extract coronary heart disease echocardiogram features has become a problem in the industry. SUMMARY

[0004] The present application provides a coronary heart disease echocardiogram pathological feature extraction method and system, which can effectively correlate and integrate the static structure and dynamic function of the heart to extract coronary heart disease echocardiogram features.

[0005] In a first aspect, the present application provides a coronary heart disease echocardiogram pathological feature extraction method, comprising the following steps: Ultrasound monitoring of the heart of the target coronary heart disease patient is performed to obtain an ultrasound video stream of the heart of the target coronary heart disease patient; From the ultrasound video stream, the static structure features and myocardial function features of the heart of the target coronary heart disease patient are extracted; According to the static structure features and the myocardial function features, spatial mapping analysis of the coronary artery blood supply area and myocardial injury state of the heart of the target coronary heart disease patient is performed to obtain a regional mapping matrix of the relationship between the coronary artery blood supply area and the myocardial injury state of the heart of the target coronary heart disease patient. According to the regional mapping matrix, the boundary sharpening of the pathological region of the heart of the target coronary heart disease patient is performed to obtain the boundary sharpening visualization of the pathological region in the heart of the target coronary heart disease patient. acquiring a myocardial motion trajectory of a target coronary heart disease patient, determining a longitudinal strain gradient of a basal segment and an apical segment of a heart of the target coronary heart disease patient according to the myocardial motion trajectory; generating a pathological feature of the heart of the target coronary heart disease patient based on the boundary sharpening of the pathological region and the longitudinal strain gradient in combination with clinical pathological data of the target coronary heart disease patient.

[0006] In some embodiments, the extracting the static structural feature and the myocardial functional feature of the heart of the target coronary heart disease patient from the ultrasound video stream specifically comprises: performing denoising processing on the ultrasound video stream to obtain a denoised ultrasound video stream; determining static data and dynamic data of the heart of the target coronary heart disease patient according to the denoised ultrasound video stream; determining the static structural feature of the heart of the target coronary heart disease patient through the static data; determining the myocardial functional feature of the heart of the target coronary heart disease patient through the dynamic data.

[0007] In some embodiments, the spatial mapping analysis of a coronary artery blood supply region and a myocardial injury state of the heart of the target coronary heart disease patient according to the static structural feature and the myocardial functional feature specifically comprises: acquiring each blood supply segment of the heart of the target coronary heart disease patient; matching the static structural feature and the myocardial functional feature according to the each blood supply segment to obtain a matching feature of a structure and a function of each blood supply segment; determining a regional mapping matrix of a relationship between the coronary artery blood supply region and the myocardial injury state of the heart of the target coronary heart disease patient according to all the matching features.

[0008] In some embodiments, the boundary sharpening of a pathological region of the heart of the target coronary heart disease patient according to the regional mapping matrix specifically comprises: determining a potential pathological region of the heart of the target coronary heart disease patient according to the regional mapping matrix; determining an edge feature of the pathological region of the heart of the target coronary heart disease patient through the potential pathological region and the ultrasound video stream; determining the boundary sharpening of the pathological region of the heart of the target coronary heart disease patient according to the edge feature.

[0009] In some embodiments, the determining the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient according to the myocardial motion trajectory specifically comprises: determining a local strain distribution map of the heart of the target coronary heart disease patient according to the myocardial motion trajectory; extracting longitudinal strain gradients of basal and apical segments of the heart of the target coronary heart disease patient from the local strain distribution map.

[0010] In some embodiments, the coronary heart disease pathology feature of the target coronary heart disease patient is generated based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradients in combination with clinical pathology data of the target coronary heart disease patient, which specifically comprises: determining geometric features of the pathological region of the heart of the target coronary heart disease patient according to the boundary sharpening visualization of the pathological region; determining functional differences of different blood supply regions of the heart of the target coronary heart disease patient according to the longitudinal strain gradients; obtaining clinical pathology data of the target coronary heart disease patient; determining coronary heart disease pathology features of the target coronary heart disease patient through the geometric features of the pathological region, the functional differences of the different blood supply regions, and the clinical pathology data.

[0011] In some embodiments, an ultrasonic diagnostic instrument equipped with a high-resolution phased array probe is selected to perform ultrasonic monitoring on the heart of the target coronary heart disease patient.

[0012] In a second aspect, the present application provides a coronary heart disease echocardiogram pathology feature extraction system, which comprises: a monitoring module configured to perform ultrasonic monitoring on the heart of the target coronary heart disease patient to obtain an ultrasonic video stream of the heart of the target coronary heart disease patient; a processing module configured to extract static structural features and myocardial functional features of the heart of the target coronary heart disease patient from the ultrasonic video stream; The processing module is further configured to perform spatial mapping analysis on coronary artery blood supply regions and myocardial injury states of the heart of the target coronary heart disease patient according to the static structural features and the myocardial functional features, to obtain a regional mapping matrix of the relationship between the coronary artery blood supply regions and the myocardial injury states of the heart of the target coronary heart disease patient, to perform boundary sharpening on the pathological region of the heart of the target coronary heart disease patient according to the regional mapping matrix, and to obtain boundary sharpening visualization of the pathological region of the heart of the target coronary heart disease patient. The processing module is further configured to obtain myocardial motion trajectories of the target coronary heart disease patient, and to determine longitudinal strain gradients of basal and apical segments of the heart of the target coronary heart disease patient according to the myocardial motion trajectories. an execution module configured to generate coronary heart disease pathology features of the target coronary heart disease patient based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradients in combination with clinical pathology data of the target coronary heart disease patient.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the pathological feature extraction method of coronary heart disease echocardiogram as described above.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the pathological feature extraction method of coronary heart disease echocardiogram as described above.

[0015] The technical scheme provided by the embodiments of the present application has the following beneficial effects: In the pathological feature extraction method and system of coronary heart disease echocardiogram provided by the present application, first, the heart of the target coronary heart disease patient is monitored by ultrasound to obtain an ultrasound video stream of the heart of the target coronary heart disease patient; static structural features and myocardial functional features of the heart of the target coronary heart disease patient are extracted from the ultrasound video stream; the spatial mapping analysis of the coronary artery blood supply area and the myocardial injury state of the heart of the target coronary heart disease patient is performed according to the static structural features and the myocardial functional features, to obtain a regional mapping matrix of the relationship between the coronary artery blood supply area and the myocardial injury state of the heart of the target coronary heart disease patient, the boundary sharpening of the pathological region of the heart of the target coronary heart disease patient is performed according to the regional mapping matrix, to obtain the boundary sharpening visualization of the pathological region in the heart of the target coronary heart disease patient; the myocardial motion trajectory of the target coronary heart disease patient is obtained, the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient is determined according to the myocardial motion trajectory; and the pathological features of the heart of the target coronary heart disease patient are generated by combining the boundary sharpening visualization of the pathological region, the longitudinal strain gradient, and the clinical pathological data of the target coronary heart disease patient.

[0016] It can be seen that, in the coronary heart disease echocardiogram feature extraction process of the present application, the dynamic images of the heart of the coronary heart disease patient are collected through real-time ultrasound video stream, the static structure and myocardial functional features are quantitatively extracted therefrom, the spatial mapping relationship matrix of the coronary artery blood supply area and the myocardial injury state is constructed, the corresponding injured myocardial region of the responsible blood vessel is accurately positioned, and the pathological region is subjected to boundary sharpening processing by using the mapping to significantly improve the visualization clarity and positioning accuracy; at the same time, the basal segment-apical segment longitudinal strain gradient is obtained by analyzing the myocardial motion trajectory, and the overall systolic coordination of the heart is quantitatively evaluated; finally, the sharpened local pathological visualization, the strain gradient reflecting the overall function, and the clinical pathological data of the patient are deeply fused to generate a comprehensive, accurate, and personalized heart pathological feature evaluation report. By using the above scheme, the static structure and dynamic function of the heart can be effectively associated and fused to extract the features of the coronary heart disease echocardiogram. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1is an exemplary flow chart of a method for extracting coronary heart disease echocardiogram pathological features according to some embodiments of the present application; Figure 2 is a structural diagram of a heart ultrasound monitoring chart according to some embodiments of the present application; Figure 3 is an exemplary flow chart of determining boundary sharpening visualization according to some embodiments of the present application; Figure 4 is a structural schematic diagram of a coronary heart disease echocardiogram pathological feature extraction system according to some embodiments of the present application; Figure 5 is a structural schematic diagram of a computer device for implementing a coronary heart disease echocardiogram pathological feature extraction method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in combination with the drawings in the specification and specific embodiments.

[0019] Reference Figure 1 The figure is an exemplary flow chart of a method for extracting coronary heart disease echocardiogram pathological features according to some embodiments of the present application, which mainly includes the following steps: In step 101, the heart of the target coronary heart disease patient is monitored by ultrasound to obtain an ultrasound video stream of the heart of the target coronary heart disease patient.

[0020] In specific implementation, an ultrasound diagnostic instrument (such as Philips series, Siemens Acuson S3000) equipped with a high-resolution phased array probe is selected to monitor the heart of the target coronary heart disease patient, and in some embodiments, reference is made to Figure 2 The figure is a structural diagram of a heart ultrasound monitoring chart according to some embodiments of the present application, and records dynamic images of 3-5 complete cardiac cycles, and the recorded dynamic images are used as the ultrasound video stream of the heart of the target coronary heart disease patient, wherein the ultrasound video stream represents the video stream of the ultrasound detection of the heart of the target coronary heart disease patient, and can be used for analyzing the heart of the target coronary heart disease patient; in other embodiments, other ways of monitoring can also be used.

[0021] It should be noted that the probe frequency is flexibly adjusted according to the patient's body type and heart depth to ensure clear imaging, and the target coronary heart disease patient is placed in a left lateral position, the probe is placed on the conventional section of the 3-5 intercostal and apical part of the left sternal margin, and if necessary, the subxiphoid section is increased to reduce the interference of lung gas, and in the parameter setting, the two-dimensional ultrasound mode is turned on, and the gain, dynamic range and frame frequency (maintained ≥60 frames / second) are reasonably adjusted.

[0022] In step 102, static structural features and myocardial functional features of the target coronary heart disease patient's heart are extracted from the ultrasound video stream.

[0023] In some embodiments, the extraction of the static structural features and the myocardial functional features of the target coronary heart disease patient's heart from the ultrasound video stream can be achieved by the following steps: The ultrasound video stream is denoised to obtain a denoised ultrasound video stream; The static data and the dynamic data of the target coronary heart disease patient's heart are determined according to the denoised ultrasound video stream; The static structural features of the target coronary heart disease patient's heart are determined by the static data; The myocardial functional features of the target coronary heart disease patient's heart are determined by the dynamic data.

[0024] In a specific implementation, first, the ultrasound video stream is denoised by using a Gaussian filtering algorithm. The Gaussian filtering algorithm performs weighted average on the neighboring pixels according to the Gaussian distribution, smoothes the image, and suppresses the Gaussian noise, thereby obtaining a denoised ultrasound video stream. Then, an edge detection algorithm is used to automatically identify the endocardial and epicardial boundaries from the denoised ultrasound video stream, and an elliptical fitting method is used to calculate the intracavity diameter and the wall thickness. The endocardial and epicardial boundaries, the intracavity diameter, and the wall thickness are taken as the static data of the target coronary heart disease patient's heart, wherein the static data represents the data of the target coronary heart disease patient's heart that does not change with time. A classical method based on feature point tracking (such as the optical flow method) is used to extract the myocardial motion in a complete cardiac cycle from the denoised ultrasound video stream, and the extracted myocardial motion is taken as the dynamic data of the target coronary heart disease patient's heart, wherein the dynamic data represents the data of the target coronary heart disease patient's heart that changes with time. Then, an image segmentation technique (such as a U-Net model based on region growing and deep learning) is used to accurately outline the contours of each structure in the heart in combination with the static data, and the contours of each structure are taken as the static structural features of the target coronary heart disease patient's heart. A speckle tracking imaging technique (such as a tracking algorithm based on the optical flow method and deep learning) is used to calculate the longitudinal and circumferential strain and strain rate, and the myocardial motion velocity of each myocardial segment in combination with the dynamic data, and the longitudinal and circumferential strain and strain rate, and the myocardial motion velocity of each myocardial segment are taken as the myocardial functional features of the target coronary heart disease patient's heart. In other embodiments, other methods can also be used to determine the static structural features and the myocardial functional features, which are not limited here.

[0025] It should be noted that the static structural features in the present application represent the features of the contour structure of the target coronary heart disease patient's heart, and the myocardial functional features represent the features of the function of the target coronary heart disease patient's heart when it is beating, which can be used for analyzing the features of the target coronary heart disease patient's heart.

[0026] In step 103, the spatial mapping analysis of the target coronary heart disease patient's heart coronary artery blood supply area and myocardial injury state is performed according to the static structural features and the myocardial functional features, a regional mapping matrix of the relationship between the target coronary heart disease patient's heart coronary artery blood supply area and myocardial injury state is obtained, and the pathological region of the target coronary heart disease patient's heart is subjected to boundary sharpening according to the regional mapping matrix, so as to obtain the boundary sharpening imaging of the pathological region in the target coronary heart disease patient's heart.

[0027] In some embodiments, the spatial mapping analysis of the target coronary heart disease patient's heart coronary artery blood supply area and myocardial injury state according to the static structural features and the myocardial functional features can be realized by the following steps: Obtaining each blood supply segment of the target coronary heart disease patient's heart; Matching the static structural features and the myocardial functional features according to the each blood supply segment to obtain the matching features of each blood supply segment structure and function; Determining the regional mapping matrix of the relationship between the target coronary heart disease patient's heart coronary artery blood supply area and myocardial injury state according to all the matching features.

[0028] It should be noted that the static structural features (such as the degree of coronary artery stenosis, plaque distribution, and vascular anatomical segmentation) and the myocardial functional features (such as myocardial segment strain, ejection fraction, and blood flow reserve) are anatomically mapped to the blood supply area in space and follow the pathological chain of "vessel obstruction-myocardial ischemia-functional injury" in time, while the coronary artery blood supply has regional specificity, and the position and degree of myocardial functional injury are highly related to the responsible vascular lesions. Therefore, the association degree of structural abnormalities (such as left anterior descending branch stenosis) and corresponding myocardial segment functional abnormalities can be quantified by combining the static structural features and the myocardial functional features.

[0029] In specific implementation, firstly, based on a 17-segment left ventricular model and the anatomical blood supply distribution patterns of the coronary arteries, the heart of the target coronary artery disease patient is divided into various blood supply segments (such as the anterior wall and anterior septal segments supplied by the left anterior descending artery). Here, the blood supply segment represents the myocardial region supplied by the coronary arteries and their branches. Next, a data alignment algorithm is used to match the static structural features corresponding to each blood supply segment with the myocardial functional features, constructing a matching feature vector for the structure and function of each blood supply segment. This matching feature vector is used as the matching feature for the corresponding blood supply segment's structure and function. Features represent the structural and functional characteristics of the blood supply segments. Finally, the Pearson correlation coefficient algorithm is used to calculate the correlation values ​​between the matching features of each coronary artery blood supply area and the corresponding blood supply segment. A matrix is ​​constructed with the coronary artery blood supply area as the row and the damage indicators of each blood supply segment (such as the strain reduction magnitude and the ventricular wall motion abnormality score) as the column. The calculated correlation values ​​are mapped to the corresponding positions of the matrix elements. The mapped matrix is ​​used as the region mapping matrix of the relationship between the coronary artery blood supply area and the myocardial injury state of the target coronary heart disease patient. Other methods can be used to determine this in other embodiments, which are not limited here.

[0030] It should be noted that the region mapping matrix in this application represents the degree of correlation between the coronary artery blood supply area and the myocardial injury state, and can be used to provide quantitative basis for the pathological analysis of the heart of the target coronary heart disease patient.

[0031] In some embodiments, reference Figure 3 As shown, this figure is an exemplary flowchart for determining boundary sharpening and imaging in some embodiments of this application. In this embodiment, the boundary sharpening of the pathological region of the heart of the target coronary heart disease patient is performed according to the region mapping matrix. The boundary sharpening and imaging of the pathological region in the heart of the target coronary heart disease patient can be achieved by the following steps: First, in step 1031, the potential pathological regions of the heart of the target coronary heart disease patient are determined according to the region mapping matrix; Secondly, in step 1032, the edge features of the cardiac pathological region of the target coronary heart disease patient are determined by the potential pathological region and the ultrasound video stream; Finally, in step 1033, the boundary of the pathological region in the heart of the target coronary heart disease patient is sharpened and visualized based on the edge features.

[0032] In a specific implementation, first, the correlation value data of the coronary artery branch and the myocardial injury index is extracted from the regional mapping matrix (for example, the absolute value of the correlation value is greater than 0.7), and the potential pathological region of the heart of the target coronary heart disease patient is located by combining the 17-segment left ventricular model and the corresponding region of each correlation value in the correlation value data, wherein the potential pathological region represents the potential pathological region of the heart of the target coronary heart disease patient; then, the edge of the potential pathological region is segmented from the ultrasound video stream based on the semantic segmentation model of deep learning, and the boundary geometry and the wall thickness gradient of the pathological region are extracted from the ultrasound video stream by combining the edge with the multi-layer convolution and the pooling operation, and the boundary geometry and the wall thickness gradient are taken as the edge features of the pathological region of the heart of the target coronary heart disease patient, wherein the edge features represent the features of the edge of the pathological region of the heart of the target coronary heart disease patient; finally, the display of the normal myocardial region is weakened and the boundary of the pathological region is strengthened by using the Gaussian blur inverse processing combined with the edge features, and the enhanced imaging image of the pathological region with clear boundary and high contrast is finally generated, which is taken as the boundary sharpening imaging of the pathological region in the heart of the target coronary heart disease patient. In other embodiments, other ways can also be used to determine, which are not limited here.

[0033] It should be noted that the boundary sharpening imaging in the present application represents the imaging image of the pathological region in the heart of the target coronary heart disease patient after enhancement, which can be used to locate the pathological region in the heart of the target coronary heart disease patient, and provide intuitive and accurate visual basis for clinical diagnosis.

[0034] In step 104, the myocardial motion trajectory of the target coronary heart disease patient is obtained, and the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient is determined according to the myocardial motion trajectory.

[0035] In a specific implementation, the myocardial motion trajectory of the target coronary heart disease patient can be obtained by the following method, that is, selecting the apical vertex and the mitral annulus of the end diastolic period as seed points, tracking the motion trajectory of each seed point in the entire cardiac cycle in the ultrasound video stream by speckle tracking imaging technology, and fusing all the motion trajectories by multi-view data fusion technology to construct the myocardial motion trajectory of the target coronary heart disease patient, wherein the myocardial motion trajectory represents the running trajectory of the myocardium of the target coronary heart disease patient; in other embodiments, other ways can also be used to obtain, which are not limited here.

[0036] In some embodiments, the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient can be determined according to the myocardial motion trajectory by the following steps: determine the local strain distribution map of the heart of the target coronary heart disease patient according to the myocardial motion trajectory; extract longitudinal strain gradient of basal segment and apical segment of the heart of the target coronary heart disease patient from the local strain distribution map.

[0037] In a specific implementation, determining the local strain distribution map of the heart of the target coronary heart disease patient according to the myocardial motion trajectory can be implemented in the following manner: performing frame-by-frame motion tracking of the acoustic speckle in the myocardium of the heart in the myocardial motion trajectory by using Lucas-Kanade (LK) optical flow algorithm, calculating displacement vectors of the myocardial speckle between adjacent frames by using Euclidean distance algorithm, taking each displacement vector as the strain at the corresponding position, and corresponding the strains at the positions to generate the local strain distribution map of the heart of the target coronary heart disease patient, wherein the local strain distribution map represents the strain distribution map of the local region of the heart of the target coronary heart disease patient, and intuitively presents the systolic and diastolic function states of the myocardium in different regions. In other embodiments, other manners can also be used for determination, which are not limited herein.

[0038] In a specific implementation, extracting the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient from the local strain distribution map can be implemented in the following manner: locating the basal segment (mitral annulus to the upper edge of the papillary muscle) and the apical segment (lower edge of the papillary muscle to the apex) of the heart of the target coronary heart disease patient from the ultrasound video stream according to the 17-segment left ventricular model, wherein the basal segment includes 6 segments of the anterior wall, the anterior septum, the inferior septum, the inferior wall, the posterior wall, and the anterolateral wall, and the apical segment includes 4 segments of the anterior wall, the anterior septum, the inferior wall, and the lateral wall, extracting all strains corresponding to the basal segment and all strains corresponding to the apical segment from the local strain distribution map, calculating the average value of all strains in each segment of the basal segment and the average value of all strains in each segment of the apical segment, respectively, obtaining the average longitudinal strain of each segment of the basal segment and the average longitudinal strain of each segment of the apical segment, and taking all the average longitudinal strains as the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient. In other embodiments, other manners can be used for extraction, which are not limited herein.

[0039] It should be noted that the longitudinal strain gradient in the present application refers to the gradient of the average longitudinal strain of the basal segment and the apical segment of the left ventricle, which can be used to provide a quantitative index for evaluating the degree of myocardial dysfunction of the coronary heart disease patient.

[0040] In step 105, the cardiac pathological characteristics of the target coronary heart disease patient are generated based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradient in combination with the clinical pathological data of the target coronary heart disease patient.

[0041] In some embodiments, generating the cardiac pathological characteristics of the target coronary heart disease patient based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradient in combination with the clinical pathological data of the target coronary heart disease patient can be implemented in the following steps: determine the geometric features of the pathological region of the heart of the target coronary heart disease patient according to the boundary sharpening visualization of the pathological region; determine the functional difference of different blood supply regions of the heart of the target coronary heart disease patient according to the longitudinal strain gradient; obtain the clinical pathological data of the target coronary heart disease patient; determine the pathological features of the heart of the target coronary heart disease patient by the geometric features of the pathological region, the functional difference of the different blood supply regions, and the clinical pathological data.

[0042] In specific implementation, a contour extraction algorithm (such as Sobel operator combined with morphological operation) is used to extract the contour of the pathological region from the boundary sharpening visualization of the pathological region, and then the geometric features of the pathological region are calculated based on the basic geometric parameters of the contour, wherein the geometric features include area, perimeter, shape factor, and boundary fractal dimension, which can be used to quantify the size and morphology of the lesion; then, according to the combination of the 17-segment left ventricular model, the left ventricle is grouped according to the coronary artery blood supply region to obtain each blood supply region, i.e., the left anterior descending branch blood supply anterior wall, anterior septum, apex segment and part of the inferior wall, the left circumflex branch blood supply lateral wall, posterior wall; the right coronary artery (RCA) blood supply inferior wall, posterior septum, the average longitudinal strain of each blood supply region is extracted from the longitudinal strain gradient, the difference value of the average longitudinal strain of each blood supply region is calculated, and all the difference values calculated are taken as the functional difference of the different blood supply regions of the heart of the target coronary heart disease patient, wherein the functional difference represents the difference degree of the different blood supply regions of the heart of the target coronary heart disease patient, and the difference value of the average longitudinal strain of each blood supply region can quantitatively reflect the myocardial contractile function heterogeneity of the ischemic region; at the same time, the clinical pathological data of the target coronary heart disease patient are obtained from the pathological database corresponding to the target coronary heart disease patient, and the clinical pathological data include coronary angiography stenosis degree, serum myocardial injury marker concentration, electrocardiogram ischemic change, etc.; finally, the geometric features of the pathological region, the functional difference data, and the clinical pathological data are normalized to eliminate the dimensional influence, the results obtained by the normalization processing are input into a support vector machine classification model, the mapping relationship between the geometric features of the pathological region, the functional difference data, and the clinical pathological data and the pathological features is trained and learned through the support vector machine classification model, and then the pathological features of the heart of the target coronary heart disease patient including lesion positioning, damage degree classification, and prognosis risk assessment (low / medium / high risk) are output through the support vector machine classification model.

[0043] It should be noted that the pathological features of the heart in the present application represent the characteristics of the pathological condition of the heart of the target coronary heart disease patient, which can be used to provide the lesion basis for the target coronary heart disease patient.

[0044] In addition, another aspect of the present application, in some embodiments, the present application provides a coronary heart disease echocardiogram pathological feature extraction system, which is described with reference toFigure 4 The figure is a schematic diagram of the structure of a coronary artery disease echocardiographic pathological feature extraction system according to some embodiments of this application. The coronary artery disease echocardiographic pathological feature extraction system 400 includes: a monitoring module 401, a processing module 402, and an execution module 403, which are described below: Monitoring module 401, in this application, is mainly used to perform ultrasound monitoring on the heart of the target coronary heart disease patient to obtain an ultrasound video stream of the heart of the target coronary heart disease patient; Processing module 402, in this application, is used to extract the static structural features and myocardial function features of the heart of the target coronary heart disease patient from the ultrasound video stream; It should be noted that the processing module 402 in this application is also used to perform spatial mapping analysis on the coronary artery blood supply area and myocardial injury state of the target coronary heart disease patient according to the static structural features and the myocardial function features, to obtain a region mapping matrix of the relationship between the coronary artery blood supply area and myocardial injury state of the target coronary heart disease patient, and to sharpen the boundary of the pathological area of ​​the heart of the target coronary heart disease patient according to the region mapping matrix, to obtain the boundary sharpening imaging of the pathological area in the heart of the target coronary heart disease patient; In addition, it should be noted that the processing module 402 in this application is also used to acquire the myocardial motion trajectory of the target coronary heart disease patient and determine the longitudinal strain gradient between the basal segment and the apical segment of the heart of the target coronary heart disease patient based on the myocardial motion trajectory. The execution module 403 in this application is mainly used to generate the cardiac pathological features of the target coronary heart disease patient based on the boundary sharpening imaging of the pathological region and the longitudinal strain gradient combined with the clinical pathological data of the target coronary heart disease patient.

[0045] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for extracting pathological features from echocardiography of coronary heart disease.

[0046] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for implementing a method for extracting pathological features from echocardiography of coronary artery disease, according to some embodiments of this application. The method for extracting pathological features from echocardiography of coronary artery disease in the above embodiments can be achieved through… Figure 5 The computer device shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0047] The processor 501 can be a general central processing unit (CPU), or an application specific integrated circuit (ASIC).

[0048] The communication bus 502 can be used to transmit information between the above-mentioned components.

[0049] The memory 503 can be a readonly memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable readonly memory (EEPROM), a compact disc readonly memory (CDROM) or other optical disk storage, a magneto-optical disk storage, a magnetic disk or other magnetic storage device, or any other medium capable of storing desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this. The memory 503 can exist independently, and is connected to the processor 501 through the communication bus 502. The memory 503 can also be integrated with the processor 501.

[0050] The memory 503 is used to store program code for implementing the scheme of the present application, and is controlled by the processor 501 to execute. The processor 501 is used to execute the program code stored in the memory 503. The program code can include one or more software modules. The method used in the above-mentioned embodiments can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0051] The communication interface 504 uses any transceiver-like device to communicate with other devices or communication networks, such as an Ethernet, a radio access network (RAN), a wireless local area network (WLAN), etc.

[0052] In a particular implementation, as one embodiment, the computer device can include multiple processors, each of which can be a single CPU processor or a multi-CPU processor. A processor herein can refer to one or more devices, circuits, and / or processing cores for processing data, such as computer program instructions.

[0053] The computer device described above can be a general purpose computer device or a special purpose computer device. In a particular implementation, the computer device can be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of the computer device.

[0054] In addition, the present application also provides a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the coronary heart disease echocardiogram pathological feature extraction method described above.

[0055] Although the preferred embodiments of the present application have been described, those skilled in the art who are informed of the basic inventive concept can make additional changes and modifications to the embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0056] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A method for extracting pathological features of coronary heart disease echocardiogram, characterized in that, The method comprises the following steps: ultrasonic monitoring is performed on a heart of a target coronary heart disease patient to obtain an ultrasonic video stream of the heart of the target coronary heart disease patient; static structural features and myocardial functional features of the heart of the target coronary heart disease patient are extracted from the ultrasonic video stream; spatial mapping analysis is performed on coronary artery blood supply regions and myocardial injury states of the heart of the target coronary heart disease patient according to the static structural features and the myocardial functional features, a regional mapping matrix of a relationship between the coronary artery blood supply regions and the myocardial injury states of the heart of the target coronary heart disease patient is obtained, and pathological regions of the heart of the target coronary heart disease patient are subjected to boundary sharpening according to the regional mapping matrix, to obtain boundary sharpening imaging of the pathological regions of the heart of the target coronary heart disease patient; myocardial motion trajectories of the target coronary heart disease patient are obtained, and longitudinal strain gradients of basal segments and apical segments of the heart of the target coronary heart disease patient are determined according to the myocardial motion trajectories; the boundary sharpening imaging of the pathological regions and the longitudinal strain gradients are combined with clinical pathological data of the target coronary heart disease patient to generate a cardiac pathological feature of the target coronary heart disease patient.

2. The method of claim 1, wherein, The extraction of the static structural features and the myocardial functional features of the heart of the target coronary heart disease patient from the ultrasonic video stream specifically comprises: noise removal processing is performed on the ultrasonic video stream to obtain a denoised ultrasonic video stream; static data and dynamic data of the heart of the target coronary heart disease patient are determined according to the denoised ultrasonic video stream; the static structural features of the heart of the target coronary heart disease patient are determined through the static data; the myocardial functional features of the heart of the target coronary heart disease patient are determined through the dynamic data.

3. The method of claim 1, wherein, The spatial mapping analysis on the coronary artery blood supply regions and the myocardial injury states of the heart of the target coronary heart disease patient according to the static structural features and the myocardial functional features specifically comprises: all blood supply segments of the heart of the target coronary heart disease patient are obtained; the static structural features and the myocardial functional features are matched according to the blood supply segments, to obtain matching features of segment structures and functions; the regional mapping matrix of the relationship between the coronary artery blood supply regions and the myocardial injury states of the heart of the target coronary heart disease patient is determined according to all the matching features.

4. The method of claim 1, wherein, The boundary sharpening of the pathological regions of the heart of the target coronary heart disease patient according to the regional mapping matrix specifically comprises: potential pathological regions of the heart of the target coronary heart disease patient are determined according to the regional mapping matrix; edge features of the pathological regions of the heart of the target coronary heart disease patient are determined through the potential pathological regions and the ultrasonic video stream; the boundary sharpening imaging of the pathological regions of the heart of the target coronary heart disease patient is determined according to the edge features.

5. The method of claim 1, wherein, The determination of the longitudinal strain gradients of the basal segments and the apical segments of the heart of the target coronary heart disease patient according to the myocardial motion trajectories specifically comprises: local strain distribution maps of the heart of the target coronary heart disease patient are determined according to the myocardial motion trajectories; the longitudinal strain gradients of the basal segments and the apical segments of the heart of the target coronary heart disease patient are extracted from the local strain distribution maps.

6. The method of claim 1, wherein, The pathological feature of the target coronary heart disease patient is generated based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradient in combination with clinical pathological data of the target coronary heart disease patient, and specifically includes: determining the geometric feature of the pathological region of the heart of the target coronary heart disease patient according to the boundary sharpening visualization of the pathological region; determining the functional difference of different blood supply regions of the heart of the target coronary heart disease patient according to the longitudinal strain gradient; acquiring the clinical pathological data of the target coronary heart disease patient; determining the pathological feature of the heart of the target coronary heart disease patient by the geometric feature of the pathological region, the functional difference of the different blood supply regions, and the clinical pathological data.

7. The method of claim 1, wherein, An ultrasonic diagnostic instrument equipped with a high-resolution phased array probe is selected to perform ultrasonic monitoring on the heart of the target coronary heart disease patient.

8. A system for extracting echocardiographic pathologies of coronary heart disease, characterized by, The method comprises: a monitoring module configured to perform ultrasonic monitoring on the heart of the target coronary heart disease patient to obtain an ultrasonic video stream of the heart of the target coronary heart disease patient; a processing module configured to extract static structural features and myocardial functional features of the heart of the target coronary heart disease patient from the ultrasonic video stream; the processing module is further configured to perform spatial mapping analysis on the coronary artery blood supply region and the myocardial injury state of the heart of the target coronary heart disease patient according to the static structural features and the myocardial functional features, to obtain a regional mapping matrix of the relationship between the coronary artery blood supply region and the myocardial injury state of the heart of the target coronary heart disease patient, to perform boundary sharpening on the pathological region of the heart of the target coronary heart disease patient according to the regional mapping matrix, and to obtain the boundary sharpening visualization of the pathological region of the heart of the target coronary heart disease patient; the processing module is further configured to acquire the myocardial motion trajectory of the target coronary heart disease patient, and to determine the longitudinal strain gradient of the basal segment and the apical segment of the heart of the target coronary heart disease patient according to the myocardial motion trajectory; an execution module configured to generate the pathological feature of the heart of the target coronary heart disease patient based on the boundary sharpening visualization of the pathological region and the longitudinal strain gradient in combination with the clinical pathological data of the target coronary heart disease patient.

9. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the coronary heart disease echocardiogram pathological feature extraction method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the coronary heart disease echocardiogram pathological feature extraction method according to any one of claims 1 to 7.

Images