Mitral valve function evaluation device
By combining echocardiography and blood flow velocity time series, and using multiple models to automatically segment and evaluate the structure and function of the mitral valve, the problem of low accuracy and efficiency in the existing technology for mitral valve function assessment is solved, and efficient and accurate mitral valve function assessment is achieved.
Patent Information
- Application Number
- CN202511484593.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-16
- Publication Date
- 2026-02-17
AI Technical Summary
Current mitral valve function assessment relies on the judgment of experienced physicians, resulting in low accuracy and efficiency. Furthermore, existing artificial intelligence methods often depend on single-modal features, leading to inaccurate assessment results.
By acquiring echocardiograms and blood flow velocity time series during the cardiac cycle of the target subject, and using mitral valve segmentation models, valve motion analysis models, hemodynamic models, and calcification detection models, the mask image is automatically segmented and mitral valve structure, blood flow, and calcification features are extracted to comprehensively evaluate mitral valve function.
This technology enables automated mitral valve function assessment without human intervention, improving the accuracy and efficiency of the assessment and solving the problems of poor accuracy and consistency in existing technologies.
Smart Images
Figure CN121533756A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of medical technology, and more specifically, to a mitral valve function assessment device. Background Technology
[0002] Mitral valve disease is one of the most common heart valve disorders, including mitral stenosis and mitral regurgitation. Its functional assessment relies on transthoracic echocardiography (TTE) and continuous-wave Doppler hemodynamic monitoring. Traditional mitral valve function assessment involves experienced physicians determining the presence of mitral valve disease based on TTE and continuous-wave Doppler hemodynamic monitoring. This traditional method is highly dependent on operator experience, resulting in low efficiency, insufficient accuracy, and significant reliance on physician subjectivity. Furthermore, existing artificial intelligence methods often rely on single-modal features (such as image segmentation or blood flow signals), leading to lower accuracy in mitral valve function assessment results. Summary of the Invention
[0003] One objective of this application is to provide a new technical solution for mitral valve function assessment, which solves the problems of low accuracy, low efficiency and poor consistency in mitral valve function assessment in related technologies, thereby improving the accuracy and efficiency of mitral valve function assessment.
[0004] According to a first aspect of this application, a mitral valve function assessment device is provided, comprising a processor and a memory, wherein the memory stores a program or instructions executable on the processor, and the program or instructions, when executed by the processor, perform the following steps: Acquire a first echocardiogram of the target object at the end of systole and a second echocardiogram at the end of diastole within the cardiac cycle, as well as a time series of blood flow velocity of the target object from the end of systole to the end of diastole within the cardiac cycle; Based on the first echocardiogram, the second echocardiogram, and the mitral valve segmentation model, determine the first mask image corresponding to the first echocardiogram and the second mask image corresponding to the second echocardiogram; Based on the first mask image and the second mask image, the mitral valve structural features of the target object are determined; Based on the region of interest in calcification in the first echocardiogram, the mitral valve calcification features of the target object are determined; wherein, the region of interest in calcification is the region in the first mask image corresponding to the leaflets and annulus of the mitral valve in the first echocardiogram; Based on the blood flow velocity time series, the mitral valve blood flow characteristics of the target object are determined; Based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features, the mitral valve function assessment results of the target object are determined and output.
[0005] Optionally, determining and outputting the mitral valve function assessment result of the target object based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features includes: The first mask image, the second mask image, and the mitral valve structural features are input into the valve motion analysis model to obtain the first evaluation result of the mitral valve function of the target object; The mitral valve blood flow characteristics are input into a hemodynamic model to obtain a second assessment result of the mitral valve function of the target object. The mitral valve calcification features are input into the calcification detection model to obtain a third evaluation result of the mitral valve function of the target object; Based on the first evaluation result, the second evaluation result, and the third evaluation result, the mitral valve function evaluation result of the target object is determined and output.
[0006] Optionally, the first assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; the second assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; and the third assessment result includes one of no or mild calcification and moderate to severe calcification. The step of determining and outputting the mitral valve function assessment result of the target object based on the first assessment result, the second assessment result, and the third assessment result includes: If the first assessment result is severe stenosis, the second assessment result is mild stenosis, and the third assessment result is severe calcification, then the mitral valve function assessment result of the target object is determined and output as severe stenosis. If the first assessment result is mild stenosis, the second assessment result is mild stenosis, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as mild stenosis; If the first assessment result is mild regurgitation, the second assessment result is mild regurgitation, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as mild regurgitation; If the first assessment result is severe regurgitation, the second assessment result is severe regurgitation, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as severe regurgitation.
[0007] Optionally, the mitral valve structural features include leaflet thickness, orifice area, annular diameter, and ellipticity. Determining the mitral valve structural features of the target object based on the first mask and the second mask includes: Based on the first mask image, determine the leaflet thickness, annular diameter, and ellipticity of the mitral valve of the target object; Based on the second mask image, the orifice area of the mitral valve of the target object is determined.
[0008] Optionally, the mitral valve calcification features include a first texture feature and a second texture feature, and determining the mitral valve calcification features of the target object based on the calcification region of interest in the first echocardiogram includes: Calculate the gray-level co-occurrence matrix of the calcified region of interest and the texture feature quantity of the gray-level co-occurrence matrix to obtain the first texture feature of the calcified region of interest; wherein, the texture feature quantity of the gray-level co-occurrence matrix includes the contrast, entropy, and second moment of the gray-level co-occurrence matrix; Calculate the local binary pattern and histogram of the local binary pattern in the calcified region of interest to obtain the second texture feature of the calcified region of interest; Based on the first texture feature and the second texture feature, the mitral valve calcification features of the target object are determined.
[0009] Optionally, the mitral valve blood flow characteristics include at least one of the following: peak regurgitation velocity, transvalvular pressure gradient, E / A ratio, effective regurgitation orifice area, peak flow velocity, and velocity-time integral.
[0010] Optionally, the mitral valve segmentation model is determined through the following steps: Obtain a first training sample set; wherein, each first training sample in the first training sample set includes a sample echocardiogram and a sample mitral valve mask image corresponding to the sample echocardiogram, wherein the pixel value of the pixel corresponding to the mitral valve leaflet tissue in the sample mitral valve mask image is a first pixel value, the pixel value of the pixel corresponding to the mitral valve annulus tissue is a second pixel value, the pixel value of the pixel corresponding to the left atrial tissue is a third pixel value, and the pixel value of the pixel corresponding to the left ventricular tissue is a fourth pixel value; The mitral valve segmentation model is trained using the first training sample set to obtain the trained mitral valve segmentation model.
[0011] Optionally, the valve motion analysis model is determined through the following steps: Obtain a second training sample set; wherein each second training sample in the second training sample set includes a first sample mask, a second sample mask, sample mitral valve structural features, and sample mitral valve functional evaluation type; The valve motion analysis model is trained using the second training sample set to obtain the trained valve motion analysis model.
[0012] Optionally, the hemodynamic model is determined through the following steps: Obtain a third training sample set; wherein each third training sample in the third training sample set includes mitral valve blood flow characteristics and mitral valve function assessment type; The hemodynamic model is trained using the third training sample set to obtain the trained hemodynamic model.
[0013] Optionally, the calcification detection model is determined through the following steps: Obtain a fourth training sample set; wherein each fourth training sample in the fourth training sample set includes mitral valve calcification features and calcification degree type of the sample; The calcification detection model is trained using the fourth training sample set to obtain the trained calcification detection model.
[0014] One beneficial effect of this application is that by determining the mitral valve function assessment result of the target object based on a first mask image, a second mask image, mitral valve structural features, mitral valve blood flow features, and mitral valve calcification features, it can solve the problem of low assessment accuracy in related technologies when performing mitral valve function assessment based on single-modal features (such as image segmentation or blood flow signals). Furthermore, by automatically segmenting the mask image and extracting mitral valve blood flow features, mitral valve structural features, and mitral valve calcification features, it can achieve automatic mitral valve function assessment without human intervention, solving the problems of low accuracy, low efficiency, and poor consistency in subjective human assessment of mitral valve function in related technologies, thus improving the accuracy and efficiency of mitral valve function assessment. Attached Figure Description
[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments of the present application and, together with their description, serve to explain the principles of the present application.
[0016] Figure 1 This is a schematic diagram of the structure of a mitral valve function assessment device provided according to an embodiment of this application; Figure 2 This is a flowchart illustrating a mitral valve function assessment method provided according to an embodiment of this application. Detailed Implementation
[0017] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present application.
[0018] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.
[0019] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.
[0020] In all the examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.
[0021] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.
[0022] Figure 1 This is a schematic block diagram of the structure of a mitral valve function assessment device 100 provided according to an embodiment of the present disclosure.
[0023] The mitral valve function assessment device 100 can be a terminal or a server to alleviate the computing power of the terminal. Furthermore, the mitral valve function assessment device 100 can be a terminal such as a portable computer, tablet computer, or PDA. The mitral valve function assessment device 100 can also be a cloud server or a network edge server.
[0024] like Figure 1 As shown, the mitral valve function assessment device 100 includes a processor 110 and a memory 120. The memory 120 stores programs or instructions that can run on the processor 110. When the program or instructions are executed by the processor 110, they achieve the following: Figure 2 The method for assessing mitral valve function is shown.
[0025] like Figure 2 As shown, it provides a method for assessing mitral valve function, including the following steps S210 to S260.
[0026] Step S210: Obtain a first echocardiogram of the target object at the end of systole and a second echocardiogram at the end of diastole within the cardiac cycle, as well as a time series of blood flow velocity of the target object from the end of systole to the end of diastole within the cardiac cycle.
[0027] In this embodiment, the cardiac cycle can refer to the time required for the heart to complete one full pumping cycle. For example, a cardiac cycle can be the time between one mitral valve closure and the next. One cardiac cycle involves the heart going through a cycle of diastole → systole → diastole.
[0028] At the end of systole, the mitral valve leaflets are completely closed, thickest, and most clearly calcified. Measuring leaflet thickness, annular diameter, ellipticity, and calcification texture at this time provides the clearest view and avoids artifacts caused by the valve orifice obstructing the view during diastole.
[0029] At the end of diastole, the mitral valve orifice is fully open, and its outline is complete, making it easy to calculate the orifice area.
[0030] Therefore, in order to improve the accuracy of the subsequent determination of the mitral valve structural features, a first echocardiogram at end-systole and a second echocardiogram at end-diastole were obtained.
[0031] Echocardiography can be either transthoracic echocardiography (TTE) or transesophageal echocardiography (TEE), without limitation here. Transthoracic echocardiography involves placing an ultrasound probe on the surface of the chest to obtain images of the heart's structure and function. Transesophageal echocardiography involves inserting an ultrasound probe into the esophagus to obtain images of the heart's structure and function. Transesophageal echocardiography produces clearer images because the ultrasound probe is closer to the heart.
[0032] The blood flow velocity time series of the target object from end-systole to end-diastole during the cardiac cycle can be obtained by continuously or high-pulse emitting ultrasound in a selected sound beam direction, performing real-time spectral analysis on the Doppler frequency shift (i.e., continuous wave Doppler signal) generated by all red blood cells along the sound beam direction, and then converting the obtained frequency-time curve into a blood flow velocity time series.
[0033] After obtaining the first and second echocardiograms in step S210, and before executing step S220, image preprocessing can be performed on the first and second echocardiograms respectively. The image preprocessing can be at least one of spatial resampling, global contrast correction, local contrast enhancement, and data augmentation, and is not limited here.
[0034] Those skilled in the art should understand that the image preprocessing described here is a common practice in the field and will not be elaborated upon here.
[0035] Step S220: Based on the first echocardiogram, the second echocardiogram, and the mitral valve segmentation model, determine the first mask image corresponding to the first echocardiogram and the second mask image corresponding to the second echocardiogram.
[0036] In this embodiment, the first echocardiogram and the second echocardiogram are segmented using a mitral valve segmentation model to obtain a first mask image corresponding to the first echocardiogram and a second mask image corresponding to the second echocardiogram.
[0037] Because the mitral valve leaflets are thin and have weak echoes, segmenting the mitral valve from a single echocardiogram mask is prone to missing edges. Therefore, it is necessary to include the larger, more clearly defined left atrium and left ventricle in the segmentation task. The mitral valve segmentation model can first locate the left atrium and left ventricle, and then locate the mitral valve leaflets and annulus. This significantly improves the segmentation accuracy of the mitral valve.
[0038] Based on this, the mitral valve segmentation model is used to perform pixel-level segmentation of the left atrium, left ventricle, mitral valve leaflets, and valve annulus in echocardiography.
[0039] In the first and second mask images, the pixel values of the pixels corresponding to the mitral valve leaflet tissue are the first pixel value, the pixel values of the pixels corresponding to the mitral valve annulus tissue are the second pixel value, the pixel values of the pixels corresponding to the left atrial tissue are the third pixel value, and the pixel values of the pixels corresponding to the left ventricular tissue are the fourth pixel value.
[0040] In one example, the mitral valve segmentation model is a spatial-channel dual-attention nnU-Net network model. The spatial-channel dual-attention nnU-Net network model is obtained by embedding a spatial attention gate (automatically generating 0–1 weights for each pixel, suppressing irrelevant background and enhancing target edges) into the encoder-decoder skip connection path of the original nnU-Net, and inserting a channel attention module (SE-Block) after the convolutional block (recalibrating the response of each channel with global information, suppressing redundant features and highlighting key anatomical structures).
[0041] In one example, the first pixel has a value of 1, the second pixel has a value of 2, the third pixel has a value of 3, and the fourth pixel has a value of 4.
[0042] In one embodiment of this application, the mitral valve segmentation model in step S220 is determined through the following steps: steps S110 to S120.
[0043] Step S110: Obtain the first training sample set.
[0044] Wherein, each first training sample in the first training sample set includes a sample echocardiogram and a sample mitral valve mask image corresponding to the sample echocardiogram. In the sample mitral valve mask image, the pixel value of the pixel corresponding to the mitral valve leaflet tissue is the first pixel value, the pixel value of the pixel corresponding to the mitral valve annulus tissue is the second pixel value, the pixel value of the pixel corresponding to the left atrial tissue is the third pixel value, and the pixel value of the pixel corresponding to the left ventricular tissue is the fourth pixel value.
[0045] In this embodiment, multiple sample echocardiograms corresponding to multiple sample objects are first obtained. Then, for each sample echocardiogram, the expert changes the pixel value of the pixel corresponding to the mitral valve leaflet in the sample echocardiogram to 1 (i.e., the first pixel value), the pixel value of the pixel corresponding to the mitral valve annulus in the sample echocardiogram to 2 (i.e., the second pixel value), the pixel value of the pixel corresponding to the left atrial tissue in the sample echocardiogram to 3 (i.e., the third pixel value), and the pixel value of the pixel corresponding to the left ventricular tissue in the sample echocardiogram to 4 (i.e., the fourth pixel value), thereby obtaining the sample mitral valve mask image corresponding to the sample echocardiogram, which serves as a first training sample. Thus, multiple first training samples corresponding to multiple sample objects can be obtained, forming a first training sample set.
[0046] It should be noted that, in order to improve the segmentation effect of the mitral valve segmentation model on echocardiograms at any time within the cardiac cycle, the sample echocardiograms obtained here are not limited to end-systole and end-diastole, but can also be echocardiograms at any other time within the cardiac cycle.
[0047] Step S120: Train the mitral valve segmentation model using the first training sample set to obtain the trained mitral valve segmentation model.
[0048] The method used to train the mitral valve segmentation model in this step is the conventional model training method, which will not be described in detail here.
[0049] Step S230: Determine the mitral valve structural features of the target object based on the first mask and the second mask.
[0050] In one embodiment of this application, the structural features of the mitral valve include leaflet thickness, valve orifice area, valve annulus diameter, and ellipticity.
[0051] In this embodiment, leaflet thickness refers to the maximum radial distance from the closure edge to the attachment edge of the mitral valve leaflet. Increased thickness indicates leaflet fibrosis or calcification, hardening the leaflet, reducing its opening range, and leading to a smaller valve orifice area. This directly results in an increased transvalvular pressure gradient, making leaflet thickness a core evaluation indicator for mitral stenosis. The valve orifice area is formed by the free edge of the leaflets during early diastole. If the orifice area is small, blood flow is obstructed, functionally manifesting as mitral stenosis, with smaller areas indicating more severe stenosis. The annular diameter is the length of the long axis of the annular plane at end-systole. A larger annular diameter results in radial traction of the leaflet occlusive edge, creating a central gap, leading to systolic blood regurgitation into the left atrium, thus becoming an important indicator for evaluating mitral regurgitation. Ellipticity is the ratio of the annular long axis to the short axis. An ellipticity approaching 1.0 (i.e., close to a circle) causes uneven circumferential tension distribution, reducing occlusive height and significantly increasing the risk of regurgitation for the same annular diameter. An ellipticity greater than 1.3 can partially preserve occlusive height, resulting in a lower probability of regurgitation.
[0052] In these embodiments, step S230, which determines the mitral valve structural features of the target object based on the first mask image and the second mask image, includes steps S2301 to S2302.
[0053] Step S2301: Based on the first mask image, determine the leaflet thickness, annular diameter, and ellipticity of the mitral valve of the target object.
[0054] In this embodiment, in the first mask image, the maximum distance is measured point by point along the normal direction of the leaflet free edge to obtain the leaflet thickness. In the first mask image, the leaflet annulus plane is fitted to form an ellipse, and the length of the major axis of the ellipse is calculated as the leaflet annulus diameter. In the fitted ellipse, the ratio of the major axis to the minor axis is calculated to obtain the ellipticity.
[0055] Those skilled in the art should understand that the methods for measuring the leaflet thickness, annular diameter, and ellipticity of the mitral valve are well-known in the field and will not be described in detail here.
[0056] Step S2302: Determine the orifice area of the mitral valve of the target object based on the second mask image.
[0057] In this embodiment, in the second mask image, the open area surrounded by the leaflet and the annulus is counted by pixels and multiplied by the pixel area to obtain the valve orifice area.
[0058] Those skilled in the art should understand that the method for measuring the orifice area of the mitral valve is well known in the field, and therefore will not be described in detail here.
[0059] Step S240: Determine the mitral valve calcification features of the target object based on the calcification region of interest in the first echocardiogram.
[0060] The calcification region of interest is the region in the first mask image corresponding to the leaflets and annulus of the mitral valve in the first echocardiogram.
[0061] In this embodiment, the image regions where the leaflets and annulus are located can be determined by the first mask image, and then the region corresponding to the pixel region in the first echocardiogram is called the region of interest for calcification.
[0062] The calcification features of the mitral valve can reflect the spatial correlation, local orderliness, and random complexity of the calcified region within the mitral valve in terms of pixel-level grayscale distribution, thereby reflecting the degree of calcification and tissue heterogeneity.
[0063] In one embodiment of this application, the mitral valve calcification features include a first texture feature and a second texture feature.
[0064] The first texture feature reflects the correlation and complexity of the overall grayscale space of the calcified region. The second texture feature reflects the pattern frequency of the pixel-level local microstructures in the calcified region.
[0065] In this embodiment, step S240, which determines the mitral valve calcification features of the target object based on the calcification region of interest in the first echocardiogram, includes steps S2401 to S2403.
[0066] Step S2401: Calculate the gray-level co-occurrence matrix of the calcified region of interest and the texture feature quantity of the gray-level co-occurrence matrix to obtain the first texture feature of the calcified region of interest.
[0067] The texture features of the gray-level co-occurrence matrix include the contrast, entropy, and second moment of the gray-level co-occurrence matrix.
[0068] In this embodiment, the joint gray-level distribution among pixels is statistically analyzed within the calcified region of interest to generate a gray-level co-occurrence matrix (GLCM). Subsequently, three statistical measures are extracted from the GLCM: contrast (reflecting the intensity of gray-level differences), entropy (reflecting the complexity / randomness of gray-level distribution), and second moment of angle (reflecting the uniformity / orderliness of gray-level distribution). These measures are used together to constitute the first texture feature to quantify the overall spatial correlation and structural complexity of the calcified region.
[0069] Step S2402: Calculate the local binary pattern of the calcified region of interest and the histogram of the local binary pattern to obtain the second texture feature of the calcified region of interest.
[0070] In this embodiment, for each pixel within the calcified region of interest, the grayscale values of its neighbors are compared with the center value using the Local Binary Pattern (LBP) operator to generate a binary code, which is then converted into a decimal LBP value. Subsequently, the distribution of LBP values for all pixels is statistically analyzed to obtain an LBP histogram. This histogram is used as the second texture feature vector to characterize the pattern frequency and fine texture distribution of the pixel-level local microstructure in the calcified region.
[0071] Step S2403: Determine the mitral valve calcification features of the target object based on the first texture feature and the second texture feature.
[0072] Step S250: Determine the mitral valve blood flow characteristics of the target object based on the blood flow velocity time series.
[0073] In one embodiment of this application, the mitral valve blood flow characteristics include at least one of peak regurgitation velocity, transvalvular pressure gradient, E / A ratio, effective regurgitation orifice area, peak flow velocity, and velocity-time integral.
[0074] In this embodiment, peak regurgitation velocity refers to the highest velocity of the regurgitation jet measured by Doppler. A higher peak regurgitation velocity indicates a larger atrioventricular pressure gradient and a more severe regurgitation. Transvalvular pressure gradient refers to the instantaneous pressure difference between the left ventricle and left atrium. Transvalvular pressure gradient reflects the intensity of valvular obstruction or regurgitation. In mitral stenosis, the transvalvular pressure gradient increases during systole; in mitral regurgitation, the transvalvular pressure gradient increases throughout systole. A larger transvalvular pressure gradient indicates a more severe stenosis or regurgitation. The E / A ratio can be defined as the ratio of the peak early diastolic velocity E to the peak atrial systolic velocity A. The E / A ratio characterizes left ventricular diastolic function; in mitral stenosis, the E peak is significantly increased, and the A peak is smaller, i.e., E / A ≫ 1. Effective regurgitant orifice area (EROA): the minimum cross-sectional area of the regurgitation jet at the valve orifice level. The effective regurgitant orifice area characterizes the degree of regurgitation. A larger EROA indicates a more severe regurgitation. A higher peak flow velocity can characterize mitral stenosis to some extent. A larger velocity-time integral can also characterize mitral stenosis to some extent.
[0075] The following describes in detail how to determine these mitral valve blood flow characteristics based on blood flow velocity time series: In the blood flow velocity time series, the negative peak value of the blood flow velocity is detected, and its absolute value is recorded as the peak regurgitation velocity. Then, according to the simplified Bernoulli equation, the blood flow velocity time series is converted point by point into an instantaneous transvalvular pressure gradient time series, and the maximum value of this series is taken as the peak transvalvular pressure gradient. The peak flow velocity E is extracted in the early diastolic phase, and the peak velocity A is automatically extracted in the late diastolic phase, and the E / A ratio is calculated. The blood flow velocity time series is numerically integrated to obtain the velocity-time integral (VTI). Based on the velocity-time integral and the peak regurgitation velocity, the effective regurgitation orifice area is determined.
[0076] Step S260: Based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features, determine and output the mitral valve function assessment results of the target object.
[0077] In one embodiment of this application, step S260 determines and outputs the mitral valve function assessment result of the target object based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features, including steps S2601 to S2604.
[0078] Step S2601: Input the first mask image, the second mask image, and the mitral valve structural features into the valve motion analysis model to obtain the first evaluation result of the mitral valve function of the target object.
[0079] In this embodiment, the first assessment result is used to characterize the function of the mitral valve, and the first assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation.
[0080] The valve motion analysis model uses the first and second mask images as geometric references, superimposing four-dimensional structural features such as leaflet thickness, orifice area, annular diameter, and ellipticity to form an eight-dimensional input of "morphology + motion". The output is a four-dimensional feature vector, representing the model's confidence score for four states (the higher the value, the higher the probability of that state).
[0081] For example, if the output four-dimensional feature vector is: mild narrowing: 0.1, severe narrowing: 0.8, mild backflow: 0.05, severe backflow: 0.05, then the probability of severe narrowing is the highest.
[0082] In one embodiment of this application, the valve motion analysis model in step S2601 is determined through the following steps S310 to S320.
[0083] Step S310: Obtain the second training sample set.
[0084] Each second training sample in the second training sample set includes a first sample mask, a second sample mask, sample mitral valve structural features, and sample mitral valve functional assessment type.
[0085] In this embodiment, the first sample mask image is obtained by segmenting the first sample echocardiogram at end-systole, and the second sample mask image is obtained by segmenting the second sample echocardiogram at end-diastole. The mitral valve function assessment type of the samples includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation.
[0086] The structural features of the mitral valve in the sample include leaflet thickness, orifice area, annular diameter, and ellipticity.
[0087] In one embodiment of this application, obtaining the second training sample set in step S310 includes: steps S310.1 to S310.5.
[0088] Step S310.1: Obtain multiple sample data corresponding to multiple sample objects to obtain a sample dataset.
[0089] The sample data of the sample object includes a first sample echocardiogram of the sample object at the end of systole and a second sample echocardiogram of the sample object at the end of diastole during the cardiac cycle, as well as the sample mitral valve function assessment type of the sample object.
[0090] In this embodiment, multiple sample data corresponding to multiple sample objects can be collected from multiple hospitals. One sample object corresponds to one sample data.
[0091] It should be noted that the first and second sample echocardiograms of the sample objects in this step correspond to the first and second echocardiograms in step S210. Therefore, the meaning of the first and second sample echocardiograms in this step will not be explained in detail here.
[0092] Step S310.2: For each sample object, based on the first sample echocardiogram, the second sample echocardiogram, and the mitral valve segmentation model of the sample object, determine the first sample mask corresponding to the first sample echocardiogram and the second sample mask corresponding to the second sample echocardiogram.
[0093] In the first sample mask and the second sample mask, the pixel value of the pixel corresponding to the leaflet tissue of the mitral valve is the first pixel value, the pixel value of the pixel corresponding to the annulus tissue of the mitral valve is the second pixel value, the pixel value of the pixel corresponding to the left atrial tissue is the third pixel value, and the pixel value of the pixel corresponding to the left ventricular tissue is the fourth pixel value.
[0094] In this embodiment, the first sample echocardiogram and the second sample echocardiogram are segmented using a mitral valve segmentation model to obtain a first sample mask corresponding to the first sample echocardiogram and a second sample mask corresponding to the second sample echocardiogram.
[0095] In one example, the first pixel has a value of 1, the second pixel has a value of 2, the third pixel has a value of 3, and the fourth pixel has a value of 4.
[0096] Step S310.3: Determine the mitral valve structural features of the sample object based on the first sample mask and the second sample mask of the sample object.
[0097] In this embodiment, the structural features of the sample mitral valve include leaflet thickness, orifice area, annular diameter, and ellipticity.
[0098] Since these structural features have already been mentioned above, they will not be elaborated upon here.
[0099] This step is basically the same as step S230 above, and will not be described again here.
[0100] In one embodiment of this application, step S310.3, which determines the mitral valve structural features of the sample object based on the first sample mask and the second sample mask of the sample object, includes steps SA1 to SA2.
[0101] Step SA1: Based on the first sample mask, determine the leaflet thickness, annular diameter, and ellipticity of the sample mitral valve of the sample object.
[0102] This step is basically the same as step S2301 above, and will not be described again here.
[0103] Step SA2: Determine the orifice area of the mitral valve of the sample object based on the second sample mask.
[0104] This step is basically the same as step S2302 above, and will not be described again here.
[0105] Step S310.4: The first sample mask, the second sample mask, the mitral valve structural features, and the mitral valve functional evaluation type of the sample object are used as a second training sample corresponding to the sample object.
[0106] Step S310.5: Obtain the second training sample set based on the multiple second training samples corresponding to the multiple sample objects.
[0107] Step S320: Train the valve motion analysis model using the second training sample set to obtain the trained valve motion analysis model.
[0108] The model training method used in this step is the same as the conventional model training method, and will not be described further here.
[0109] Step S2602: Input the mitral valve blood flow characteristics into the hemodynamic model to obtain a second evaluation result of the mitral valve function of the target object.
[0110] In this embodiment, the second assessment result is used to characterize the function of the mitral valve, and the second assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation.
[0111] The hemodynamic model takes a six-dimensional parameter (i.e., peak regurgitation velocity, transvalvular pressure gradient, E / A ratio, effective regurgitation orifice area, peak flow velocity, and velocity-time integral) as input and outputs a four-dimensional feature vector.
[0112] For example, if the hemodynamic model outputs a 4-dimensional feature vector: mild stenosis: 0.7, severe stenosis: 0.2, mild regurgitation: 0.1, severe regurgitation: 0.0, then it indicates that the probability of mild stenosis is relatively high.
[0113] In one embodiment of this application, the hemodynamic model is determined through the following steps S410 to S420.
[0114] Step S410: Obtain the third training sample set.
[0115] Each third training sample in the third training sample set includes mitral valve blood flow characteristics and mitral valve function assessment type.
[0116] In this embodiment, the mitral valve blood flow characteristics of the sample include at least one of the following: peak regurgitation velocity, transvalvular pressure gradient, E / A ratio, effective regurgitation orifice area, peak flow velocity, and velocity-time integral.
[0117] The mitral valve function assessment types included one of the following: mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation.
[0118] When acquiring the third training sample set: First, the time series of blood flow velocities for multiple sample objects from end-systole to end-diastole during the cardiac cycle are acquired. Then, for each sample object, the mitral valve blood flow characteristics are determined based on the corresponding blood flow velocity time series, thus obtaining multiple mitral valve blood flow characteristics for multiple sample objects. Finally, the mitral valve blood flow characteristics and mitral valve function assessment type corresponding to a sample object are used as a third training sample to obtain the third training sample set corresponding to multiple sample objects.
[0119] The method for determining the mitral valve blood flow characteristics in this step can be referred to in step S250 above, and will not be repeated here.
[0120] Step S420: Train the hemodynamic model using the third training sample set to obtain the trained hemodynamic model.
[0121] The model training method used in this step is the conventional model training method, which will not be elaborated on here.
[0122] Step S2603: Input the mitral valve calcification features into the calcification detection model to obtain the third evaluation result of the mitral valve function of the target object.
[0123] In this embodiment, the third assessment result is used to characterize the degree of mitral valve calcification. The third assessment result includes one of no or mild calcification, or moderate to severe calcification.
[0124] The calcification detection model outputs a 2D feature vector, representing the severity of calcification.
[0125] For example, if the two-dimensional feature vector output by the calcification detection model is: no or mild calcification: 0.2, moderate to severe calcification: 0.8, then the model has detected severe calcification.
[0126] In one embodiment of this application, the calcification detection model is determined through the following steps S510 to S520.
[0127] Step S510: Obtain the fourth training sample set.
[0128] Each fourth training sample in the fourth training sample set includes mitral valve calcification features and calcification degree type.
[0129] In this embodiment, the sample calcification level includes one of no or mild calcification, or moderate to severe calcification. The mitral valve calcification features of the sample include first sample texture features and second sample texture features.
[0130] The acquisition of the fourth training sample set includes: firstly, acquiring multiple first-sample echocardiograms of multiple sample objects at the end of systole within the cardiac cycle. Then, for each sample object, determining the sample mitral valve calcification features based on the calcified region of interest (ROI) in the corresponding first-sample echocardiogram. Specifically, determining the sample mitral valve calcification features based on the calcified ROI in the corresponding first-sample echocardiogram includes: calculating the gray-level co-occurrence matrix and the texture feature quantity of the gray-level co-occurrence matrix of the calcified ROI to obtain the first-sample texture features of the calcified ROI; and calculating the local binary pattern and the histogram of the local binary pattern of the calcified ROI to obtain the second-sample texture features of the calcified ROI.
[0131] In this step, the gray-level co-occurrence matrix and the texture feature quantity of the gray-level co-occurrence matrix of the calcified region of interest are calculated to obtain the first sample texture feature of the calcified region of interest, which can be referred to in step S2401 above. In this step, the local binary pattern and the histogram of the local binary pattern of the calcified region of interest are calculated to obtain the second sample texture feature of the calcified region of interest, which can be referred to in step S2402 above.
[0132] Step S520: Train the calcification detection model using the fourth training sample set to obtain the trained calcification detection model.
[0133] The model training method used in this step is the conventional model training method, which will not be elaborated on here.
[0134] Step S2604: Based on the first evaluation result, the second evaluation result, and the third evaluation result, determine and output the mitral valve function evaluation result of the target object.
[0135] In this embodiment, the first evaluation result, the second evaluation result, and the third evaluation result are input into the meta-learner (MLP) to obtain the mitral valve function evaluation result of the target object.
[0136] The weights and activation functions inside the meta-learner MLP can capture the complex, non-linear relationships between the first, second, and third evaluation results, thereby outputting the final mitral valve function evaluation result for the target object.
[0137] The mitral valve function assessment results include one of the following: mild stenosis, severe stenosis, mild regurgitation, or severe regurgitation.
[0138] In one embodiment of this application, the first assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; the second assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; and the third assessment result includes one of no or mild calcification and moderate to severe calcification. Step S2604 determines and outputs the mitral valve function assessment result of the target object based on the first assessment result, the second assessment result, and the third assessment result, including steps S2604.1 to S2604.4.
[0139] Step S2604.1: If the first assessment result is severe stenosis, the second assessment result is mild stenosis, and the third assessment result is severe calcification, determine and output the mitral valve function assessment result of the target object as severe stenosis.
[0140] In this embodiment, when the first assessment result output by the valve motion analysis model is severe stenosis, while the second assessment result output by the hemodynamic model is mild stenosis, the root cause of the discrepancy between these two assessment results is often severe calcification. That is, a large amount of calcium salt deposition hardens the valve leaflets, significantly reduces the opening amplitude, and slows down the flow rate, leading the valve motion analysis model to judge it as severe stenosis. However, the strong reflection and attenuation of ultrasound by calcification weakens the Doppler signal intensity, resulting in an underestimation of the peak flow velocity, causing the hemodynamic model to output mild stenosis. Therefore, when the first assessment result is severe stenosis, the second assessment result is mild stenosis, and the third assessment result is severe calcification, the meta-learner should consider the hemodynamic model's incorrect estimation of mild stenosis as being caused by calcification artifacts, thereby locking the final assessment result as severe stenosis.
[0141] Step S2604.2: If the first assessment result is mild stenosis, the second assessment result is mild stenosis, and the third assessment result is no or mild calcification, determine and output the mitral valve function assessment result of the target object as mild stenosis.
[0142] In this embodiment, when the valve motion analysis model outputs mild stenosis, the hemodynamic model outputs mild stenosis, and the calcification detection model outputs no or mild calcification, the hemodynamic model will not underestimate the degree of stenosis due to acoustic attenuation or hardening caused by calcification. That is, the output results of the hemodynamic model are highly reliable. Therefore, the final mitral valve function assessment result is determined to be mild stenosis.
[0143] Step S2604.3: If the first evaluation result is mild regurgitation, the second evaluation result is mild regurgitation, and the third evaluation result is no or mild calcification, determine and output the mitral valve function evaluation result of the target object as mild regurgitation.
[0144] In this embodiment, when the valve motion analysis model outputs severe regurgitation, the hemodynamic model outputs mild regurgitation, and the calcification detection model outputs no or mild calcification, the three evaluation results are consistent and have a slight load. Furthermore, the absence or mild calcification eliminates interference from calcification obstructing closure or acoustic shadowing causing overestimation of flow velocity, thus increasing the reliability of the hemodynamic model's output. Therefore, the final mitral valve function evaluation result output by the meta-learner is mild regurgitation.
[0145] Step S2604.4: If the first evaluation result is severe regurgitation, the second evaluation result is severe regurgitation, and the third evaluation result is no or mild calcification, determine and output the mitral valve function evaluation result of the target object as severe regurgitation.
[0146] In this embodiment, when the valve motion analysis model outputs severe regurgitation, the hemodynamic model outputs severe regurgitation, and the calcification detection model outputs no or mild calcification, the three evaluation results form a consistent peak in the regurgitation dimension. Furthermore, the absence or mild calcification eliminates interference from calcification obstructing closure or acoustic shadowing causing overestimation of flow velocity, thus increasing the reliability of the hemodynamic model's output. Therefore, the meta-learner does not need to shift towards the stenosis side to determine the final mitral valve function assessment result as severe regurgitation.
[0147] In one embodiment of this application, the mitral valve function assessment device 100 further includes a display ( Figure 1 (Not shown in the image), the display is used to show the mitral valve function assessment results of the target object.
[0148] By determining the mitral valve function assessment results for the target object based on a first mask image, a second mask image, mitral valve structural features, mitral valve blood flow features, and mitral valve calcification features, this approach addresses the low accuracy issue in related technologies that rely on single-modal features (such as image segmentation or blood flow signals) for mitral valve function assessment. Furthermore, by automatically segmenting the mask image and extracting mitral valve blood flow features, structural features, and calcification features, automated mitral valve function assessment without human intervention can be achieved. This solves the problems of low accuracy, inefficiency, and poor consistency associated with subjective human assessment of mitral valve function in related technologies, thus improving the accuracy and efficiency of mitral valve function assessment.
[0149] This application may be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this application.
[0150] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.
[0151] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.
[0152] The computer program instructions used to perform the operations of this application may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuits, such as programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), are personalized by utilizing the status information of the computer-readable program instructions. These electronic circuits can execute the computer-readable program instructions to implement various aspects of this application.
[0153] Various aspects of this application are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.
[0154] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.
[0155] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.
[0156] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It will be well known to those skilled in the art that implementation in hardware, implementation in software, and implementation using a combination of software and hardware are equivalent.
[0157] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical applications, or technological improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of this application is defined by the appended claims.
Claims
1. A mitral valve function assessment device, characterized in that, It includes a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions being executed by the processor to perform the following steps: Acquire a first echocardiogram of the target object at the end of systole and a second echocardiogram at the end of diastole within the cardiac cycle, as well as a time series of blood flow velocity of the target object from the end of systole to the end of diastole within the cardiac cycle; Based on the first echocardiogram, the second echocardiogram, and the mitral valve segmentation model, determine the first mask image corresponding to the first echocardiogram and the second mask image corresponding to the second echocardiogram; Based on the first mask image and the second mask image, the mitral valve structural features of the target object are determined; Based on the region of interest in calcification in the first echocardiogram, the mitral valve calcification features of the target object are determined; wherein, the region of interest in calcification is the region in the first mask image corresponding to the leaflet and annulus of the mitral valve in the first echocardiogram; Based on the blood flow velocity time series, the mitral valve blood flow characteristics of the target object are determined; Based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features, the mitral valve function assessment results of the target object are determined and output.
2. The apparatus according to claim 1, characterized in that, The step of determining and outputting the mitral valve function assessment result of the target object based on the first mask image, the second mask image, the mitral valve structural features, the mitral valve blood flow features, and the mitral valve calcification features includes: The first mask image, the second mask image, and the mitral valve structural features are input into the valve motion analysis model to obtain the first evaluation result of the mitral valve function of the target object; The mitral valve blood flow characteristics are input into a hemodynamic model to obtain a second assessment result of the mitral valve function of the target object. The mitral valve calcification features are input into the calcification detection model to obtain a third evaluation result of the mitral valve function of the target object; Based on the first evaluation result, the second evaluation result, and the third evaluation result, the mitral valve function evaluation result of the target object is determined and output.
3. The apparatus according to claim 2, characterized in that, The first assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; the second assessment result includes one of mild stenosis, severe stenosis, mild regurgitation, and severe regurgitation; the third assessment result includes one of no or mild calcification and moderate to severe calcification; the step of determining and outputting the mitral valve function assessment result of the target object based on the first assessment result, the second assessment result, and the third assessment result includes: If the first assessment result is severe stenosis, the second assessment result is mild stenosis, and the third assessment result is severe calcification, then the mitral valve function assessment result of the target object is determined and output as severe stenosis. If the first assessment result is mild stenosis, the second assessment result is mild stenosis, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as mild stenosis; If the first assessment result is mild regurgitation, the second assessment result is mild regurgitation, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as mild regurgitation; If the first assessment result is severe regurgitation, the second assessment result is severe regurgitation, and the third assessment result is no or mild calcification, the mitral valve function assessment result of the target object is determined and output as severe regurgitation.
4. The apparatus according to claim 1, characterized in that, The mitral valve structural features include leaflet thickness, orifice area, annular diameter, and ellipticity. Determining the mitral valve structural features of the target object based on the first and second mask images includes: Based on the first mask image, determine the leaflet thickness, annular diameter, and ellipticity of the mitral valve of the target object; Based on the second mask image, the orifice area of the mitral valve of the target object is determined.
5. The apparatus according to claim 1, characterized in that, The mitral valve calcification features include a first texture feature and a second texture feature. Determining the mitral valve calcification features of the target object based on the region of interest in calcification in the first echocardiogram includes: Calculate the gray-level co-occurrence matrix of the calcified region of interest and the texture feature quantity of the gray-level co-occurrence matrix to obtain the first texture feature of the calcified region of interest; wherein, the texture feature quantity of the gray-level co-occurrence matrix includes the contrast, entropy, and second moment of the gray-level co-occurrence matrix; Calculate the local binary pattern and histogram of the local binary pattern in the calcified region of interest to obtain the second texture feature of the calcified region of interest; Based on the first texture feature and the second texture feature, the mitral valve calcification features of the target object are determined.
6. The apparatus according to claim 1, characterized in that, The mitral valve blood flow characteristics include at least one of the following: peak regurgitation velocity, transvalvular pressure gradient, E / A ratio, effective regurgitation orifice area, peak flow velocity, and velocity-time integral.
7. The apparatus according to claim 1, characterized in that, The mitral valve segmentation model is determined through the following steps: Obtain a first training sample set; wherein, each first training sample in the first training sample set includes a sample echocardiogram and a sample mitral valve mask image corresponding to the sample echocardiogram, wherein the pixel value of the pixel corresponding to the mitral valve leaflet tissue in the sample mitral valve mask image is a first pixel value, the pixel value of the pixel corresponding to the mitral valve annulus tissue is a second pixel value, the pixel value of the pixel corresponding to the left atrial tissue is a third pixel value, and the pixel value of the pixel corresponding to the left ventricular tissue is a fourth pixel value; The mitral valve segmentation model is trained using the first training sample set to obtain the trained mitral valve segmentation model.
8. The apparatus according to claim 2, characterized in that, The valve motion analysis model was determined through the following steps: Obtain a second training sample set; wherein each second training sample in the second training sample set includes a first sample mask, a second sample mask, sample mitral valve structural features, and sample mitral valve functional evaluation type; The valve motion analysis model is trained using the second training sample set to obtain the trained valve motion analysis model.
9. The apparatus according to claim 2, characterized in that, The hemodynamic model was determined through the following steps: Obtain a third training sample set; wherein each third training sample in the third training sample set includes mitral valve blood flow characteristics and mitral valve function assessment type; The hemodynamic model is trained using the third training sample set to obtain the trained hemodynamic model.
10. The apparatus according to claim 2, characterized in that, The calcification detection model was determined through the following steps: Obtain a fourth training sample set; wherein each fourth training sample in the fourth training sample set includes mitral valve calcification features and calcification degree type of the sample; The calcification detection model is trained using the fourth training sample set to obtain the trained calcification detection model.