Mitral Valve Opening Distance Detection Method, Electronic Device, and Storage Medium

Abstract

The present invention provides a method for detecting the mitral valve opening distance, an electronic device, and a storage medium. The detection method includes obtaining the position information of the mitral valve region of interest according to the acquired current-frame cardiac image; segmenting the mitral valve region of interest corresponding to the current-frame cardiac image by using a leaflet segmentation model to obtain a mitral valve leaflet mask image; performing connected component analysis on the mitral valve leaflet mask image, and obtaining the mitral valve opening distance corresponding to the current-frame cardiac image according to the analysis result of the connected components. The present invention can automatically detect the mitral valve opening distance, provide an evaluation basis for the diagnosis of whether the mitral valve is stenotic, not only improve the overall algorithm accuracy, but also reduce the differential problems that may be caused by human factors.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a method for detecting the mitral valve opening distance, an electronic device, and a storage medium. Background Art

[0002] Mitral stenosis is the most common disease in valvular heart disease and its incidence is increasing year by year. It is mainly seen in rheumatic heart disease, congenital malformations, and the elderly. Due to various reasons, the structure of the mitral valve of the heart changes, resulting in a smaller opening amplitude, limited opening, or obstruction of the mitral valve, causing a series of abnormal changes in the structure and function of the heart, such as obstruction of blood flow in the left atrium and reduction of the blood volume returning to the left ventricle. At present, there is no innovative progress in the diagnosis process of this disease in the medical field. Therefore, in-depth analysis of the abnormal structure of the mitral valve using medical images is of great significance for the prevention and diagnosis of valvular heart disease.

[0003] With the continuous improvement of software technology and hardware performance, computer-aided diagnosis technology is increasingly widely used in the medical field. By visualizing the long-axis leaflets of the human mitral valve and providing the opening distance and dynamic spectrogram at the narrowest point of the leaflet tip, it can help physicians obtain more diagnostic information. Therefore, the measurement of the opening distance at the narrowest point of the long-axis leaflets of the mitral valve is particularly important.

[0004] It should be noted that the information disclosed in the background art of this invention is only intended to deepen the understanding of the general background art of the present invention, and should not be regarded as an admission or any form of suggestion that this information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for detecting the mitral valve opening distance, an electronic device, and a storage medium, which can automatically detect the opening distance at the narrowest point of the long-axis leaflets of the mitral valve, provide a basis for the diagnosis of whether the mitral valve is stenotic, not only can improve the overall algorithm accuracy, but also can reduce the differential problems that may be caused by human factors, and thus can better assist doctors in improving the diagnosis efficiency.

[0006] To achieve the above purpose, the present invention provides a method for detecting the mitral valve opening distance, including:

[0007] Obtain the position information of the region of interest of the mitral valve according to the acquired current-frame cardiac image;

[0008] Segment the region of interest of the mitral valve corresponding to the current-frame cardiac image by using a leaflet segmentation model to obtain a mitral valve leaflet mask image;

[0009] Perform connected component analysis on the mitral valve leaflet mask image. If the analysis result of the connected component is that there are two connected components with pixel areas greater than the first preset threshold in the mitral valve leaflet mask image, then extract two leaflet contours based on these two connected components with pixel areas greater than the first preset threshold, and obtain the mitral valve opening distance corresponding to the current frame of the cardiac image according to the coordinates of each pixel point on the two leaflet contours.

[0010] Optionally, the obtaining the mitral valve opening distance corresponding to the current frame of the cardiac image according to the coordinates of each pixel point on the two leaflet contours includes:

[0011] Calculate the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours:

[0012] Obtain the mitral valve opening distance corresponding to the current frame of the cardiac image according to the minimum pixel distance and the pre-obtained corresponding relationship between the pixel distance and the physical distance.

[0013] Optionally, the calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours includes:

[0014] Step A: Determine a pixel point on each of the two leaflet contours as a starting point according to the coordinates of each pixel point on the two leaflet contours;

[0015] Step B: Draw a circle with the two determined starting points as the two endpoints of a diameter;

[0016] Step C: Determine whether the drawn circle has new intersections with the two leaflet contours. If not, execute Step D; if so, execute Step E;

[0017] Step D: Take the diameter of the drawn circle as the minimum pixel distance between the two leaflet contours;

[0018] Step E: Determine whether the number of the new intersections is greater than or equal to 2. If so, execute Step E1; if not, execute Step E2;

[0019] Step E1: Draw a circle with two of the new intersections on different leaflet contours as the two endpoints of a new diameter, and return to execute Step C;

[0020] Step E2: Draw a circle with the new intersection and the original intersection on the other leaflet contour as the two endpoints of a new diameter, and return to execute Step C.

[0021] Optionally, for Step E1, the method further includes:

[0022] Determine whether the absolute value of the difference between the distance between the two new intersection points located on different leaflet contours and the diameter of the currently drawn circle is less than a second preset threshold;

[0023] If so, take the distance between the two new intersection points as the minimum pixel distance between the two leaflet contours;

[0024] If not, draw a circle with the two new intersection points as the two endpoints of a new diameter;

[0025] For step E2, the method further includes:

[0026] Determine whether the absolute value of the difference between the distance between the new intersection point and the original intersection point located on another leaflet contour and the diameter of the currently drawn circle is less than the second preset threshold;

[0027] If so, take the distance between the new intersection point and the original intersection point located on another leaflet contour as the minimum pixel distance between the two leaflet contours;

[0028] If not, draw a circle with the new intersection point and the original intersection point located on another leaflet contour as the two endpoints of a new diameter.

[0029] Optionally, determining a pixel point as a starting point on each of the two leaflet contours respectively according to the coordinates of each pixel point on the two leaflet contours includes:

[0030] For each leaflet contour, according to the coordinates of each pixel point on the leaflet contour, take the pixel point located on the leftmost side as the starting point on the leaflet contour; or

[0031] For each leaflet contour, according to the coordinates of each pixel point on the leaflet contour, take the pixel point with the smallest sum of the X coordinate and the Y coordinate as the starting point on the leaflet contour.

[0032] Optionally, calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours includes:

[0033] According to the coordinates of each pixel point on the two leaflet contours, use a support vector machine to determine the decision boundary for distinguishing the two leaflet contours and the support vectors located on each of the leaflet contours, and the support vector is the pixel point on the leaflet contour that is closest to the decision boundary;

[0034] For each support vector, draw a perpendicular line from the support vector to the decision boundary. If the perpendicular line intersects with the other leaflet contour, take the support vector as the target support vector;

[0035] For each of the target support vectors, calculate the pixel distance between the target support vector and the corresponding intersection point;

[0036] Take the calculated minimum pixel distance as the minimum pixel distance between the two leaflet contours.

[0037] Optionally, the determining the decision vectors and the support vectors on each of the leaflet contours includes:

[0038] Perform binary classification on each pixel point on the two leaflet contours to obtain the pixel points on the upper leaflet contour and the pixel points on the lower leaflet contour, so as to determine the decision boundary;

[0039] Take the pixel point on the upper leaflet contour that is closest to the decision boundary as the support vector on the upper leaflet contour, and take the pixel point on the lower leaflet contour that is closest to the decision boundary as the support vector on the lower leaflet contour.

[0040] Optionally, after obtaining the mitral valve opening distances corresponding to all frames of cardiac motion images, the method further includes:

[0041] Draw the two leaflet contours, the mitral valve opening distance diameter line, and the text content of the maximum mitral valve opening distance on the cardiac motion image corresponding to the maximum mitral valve opening distance and output.

[0042] Optionally, the method further includes:

[0043] Draw an opening distance spectrogram for characterizing the corresponding relationship between the number of frames and the mitral valve opening distance according to the time sequence and the mitral valve opening distance corresponding to each frame of cardiac motion image.

[0044] Optionally, obtaining the position information of the mitral valve region of interest according to the current frame of cardiac motion image includes:

[0045] Use an object detection model to detect the current frame of cardiac motion image to obtain the position information of the mitral valve region of interest.

[0046] To achieve the above object, the present invention also provides an electronic device, including a processor and a memory, where a computer program is stored on the memory, and when the computer program is executed by the processor, the mitral valve opening distance detection method described above is implemented.

[0047] To achieve the above object, the present invention also provides a readable storage medium, where a computer program is stored in the readable storage medium, and when the computer program is executed by a processor, the mitral valve opening distance detection method described above is implemented.

[0048] Compared with the prior art, the mitral valve opening distance detection method, electronic device, and storage medium provided by the present invention have the following advantages:

[0049] In the present invention, first, according to the acquired current-frame cardiac image, the position information of the mitral valve region of interest is obtained; then, according to the position information of the mitral valve region of interest, a leaflet segmentation model is used to segment the mitral valve region of interest corresponding to the current-frame cardiac image to obtain a mitral valve leaflet mask image; finally, connected component analysis is performed on the mitral valve leaflet mask image, and when the analysis result of the connected components is that there are two connected components in the mitral valve leaflet mask image with pixel areas greater than a first preset threshold, two leaflet contours are extracted according to these two connected components with pixel areas greater than the first preset threshold, and according to the coordinates of each pixel point on the two leaflet contours, the mitral valve opening distance corresponding to the current-frame cardiac image is obtained. Thus, it can be seen that the present invention can automatically detect the mitral valve opening distance, providing an evaluation basis for the diagnosis of mitral valve stenosis. It can not only improve the overall algorithm accuracy but also reduce the differential problems that may be caused by human factors, and further can better assist doctors in improving the diagnosis efficiency and effectively reducing the risk of misdiagnosis in the process of analyzing mitral valve abnormalities using echocardiography in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a schematic flowchart of a mitral valve opening distance detection method provided by an embodiment of the present invention;

[0051] Figure 2 is a schematic diagram of a candidate mitral valve region of interest and a final mitral valve region of interest in a cardiac image provided by a specific example of the present invention;

[0052] Figure 3 is a schematic diagram of the annotation of the mitral valve region of interest in a cardiac image provided by a specific example of the present invention;

[0053] Figure 4 For Figure 3 is a mitral valve leaflet mask image obtained by segmenting the mitral valve region of interest in the shown cardiac image;

[0054] Figure 5 is a schematic diagram of a doctor outlining the mitral valve opening distance diameter line provided by a specific example of the present invention;

[0055] Figure 6 is a schematic diagram of the drawing of the border of the distance diameter line provided by a specific example of the present invention;

[0056] Figure 7 is a schematic flowchart of a specific process for calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention;

[0057] Figure 8 Schematic diagram of the principle for calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention;

[0058] Figure 9 Schematic diagram of the result for calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention;

[0059] Figure 10 Schematic diagram of the iterative process for calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention;

[0060] Figure 11 Schematic diagram of the specific process for calculating the minimum pixel distance between two leaflet contours provided by the second embodiment of the present invention;

[0061] Figure 12 Schematic diagram of the principle for calculating the minimum pixel distance between two leaflet contours provided by the second embodiment of the present invention;

[0062] Figure 13 Spectrum diagram of the opening spacing provided by a specific example of the present invention;

[0063] Figure 14 Schematic diagram for drawing the cardiac image corresponding to the maximum mitral valve opening spacing provided by a specific example of the present invention;

[0064] Figure 15 Schematic diagram of the structure of the target detection model provided by a specific example of the present invention;

[0065] Figure 16a Schematic diagram of the structure of the first residual module provided by a specific example of the present invention;

[0066] Figure 16b Schematic diagram of the structure of the second residual module provided by a specific example of the present invention;

[0067] Figure 16c Schematic diagram of the structure of the third residual module provided by a specific example of the present invention;

[0068] Figure 16d Schematic diagram of the structure of the fourth residual module provided by a specific example of the present invention;

[0069] Figure 17 Annotated sample cardiac image provided by a specific example of the present invention;

[0070] Figure 18 Schematic diagram of the structure of the dense connection block provided by a specific example of the present invention;

[0071] Figure 19 Structural schematic diagram of a leaflet segmentation model provided for a specific example of the present invention;

[0072] Figure 20 Structural schematic diagram of a bottleneck layer provided for a specific example of the present invention;

[0073] Figure 21 Structural schematic diagram of a transition block provided for a specific example of the present invention;

[0074] Figure 22 Structural schematic diagram of an upward transition block provided for a specific example of the present invention;

[0075] Figure 23 Block diagram structural schematic diagram of an electronic device provided for an embodiment of the present invention.

[0076] Among them, the reference numerals are as follows:

[0077] Processor - 101; Communication interface - 102; Memory - 103; Communication bus - 104. Specific embodiments

[0078] The following further elaborates on the mitral valve opening distance detection method, electronic device, and storage medium proposed by the present invention in conjunction with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the accompanying drawings are in a very simplified form and all use non-precise scales, only for the purpose of conveniently and clearly assisting in explaining the embodiments of the present invention. In order to make the purpose, features, and advantages of the present invention more obvious and understandable, please refer to the accompanying drawings. It should be known that the structures, scales, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention. Any modification of the structure, change in the proportional relationship, or adjustment of the size, in the case of being the same or similar to the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed by the present invention.

[0079] It should be noted that in this text, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element, and the term "plurality" includes the case of two.

[0080] In addition, in the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0081] The core idea of the present invention is to provide a method for detecting the mitral valve opening distance, an electronic device and a storage medium, which can automatically detect the opening distance at the narrowest part of the mitral valve long axis leaflet, provide a basis for the diagnosis of mitral valve stenosis, not only improve the overall algorithm accuracy, but also reduce the differential problems that may be caused by human factors, and thus can better assist doctors in improving the diagnosis efficiency.

[0082] It should be noted that the method for detecting the mitral valve opening distance according to the embodiment of the present invention can be applied to the electronic device according to the embodiment of the present invention, wherein the electronic device can be a personal computer, a mobile terminal, etc., and the mobile terminal can be a hardware device such as a mobile phone, a tablet computer, etc. with various operating systems.

[0083] To achieve the above idea, the present invention provides a method for detecting the mitral valve opening distance. Please refer to Figure 1 , which schematically shows a flowchart of the method for detecting the mitral valve opening distance provided by an embodiment of the present invention. As Figure 1 shown, the method for detecting the mitral valve opening distance includes the following steps:

[0084] Step S100: Obtain the position information of the mitral valve region of interest based on the acquired current-frame cardiac image.

[0085] Step S200: According to the position information of the mitral valve region of interest, use a leaflet segmentation model to segment the mitral valve region of interest corresponding to the current-frame cardiac image, so as to obtain a mitral valve leaflet mask image.

[0086] Step S300: Perform connected component analysis on the mitral valve leaflet mask image. If the analysis result of the connected components is that there are two connected components in the mitral valve leaflet mask image with pixel areas greater than a first preset threshold, then extract two leaflet contours according to these two connected components with pixel areas greater than the first preset threshold, and obtain the mitral valve opening distance corresponding to the current-frame cardiac image according to the coordinates of each pixel point on the two leaflet contours.

[0087] It can be seen that the present invention can automatically detect the mitral valve opening distance, provide an evaluation basis for the diagnosis of mitral valve stenosis, not only can improve the overall algorithm accuracy, but also can reduce the differential problems that may be caused by human factors, and further can better assist doctors to improve the diagnosis efficiency and effectively reduce the risk of misdiagnosis in the process of analyzing mitral valve abnormalities using echocardiography in the prior art.

[0088] Specifically, the current-frame cardiac image is extracted from the acquired echocardiography video (containing several cardiac cycles). The resolution of the echocardiography video can be set according to specific circumstances, such as 600×800. The echocardiography video is specifically the psax-av section image collected by an ultrasound device. It should be noted that as can be understood by those skilled in the art, the current frame is dynamically changing, that is, the current-frame cardiac image changes with time. After detecting the mitral valve opening distance of the current-frame cardiac image, continue to extract the next frame of cardiac image as the new current-frame cardiac image to continue the detection of the mitral valve opening distance until the detection of the mitral valve opening distance of all frames of cardiac images is completed. In addition, it should be noted that although the present invention is described by taking echocardiography as an example, as can be understood by those skilled in the art, the cardiac image can also be a cardiac image collected by other medical devices (such as a cardiac endoscope) other than an ultrasound device, and the present invention does not limit this.

[0089] It should be noted that, as can be understood by those skilled in the art, when the acquisition moment of the current frame cardiac image is at the time of mitral valve closure, there is only one connected domain with a pixel area greater than the first preset threshold in the mitral valve leaflet mask image corresponding to the current frame cardiac image. Since the opening distance is 0 when the mitral valve is closed, if the analysis result of the mitral valve leaflet mask image is that there is only one connected domain with a pixel area greater than the first preset threshold, it indicates that the mitral valve opening distance corresponding to the current frame cardiac image is 0. When the acquisition moment of the current frame cardiac image is at the time of mitral valve opening, there are two connected domains with a pixel area greater than the first preset threshold in the mitral valve leaflet mask image corresponding to the current frame cardiac image. The regions defined by these two connected domains are the two leaflets of the mitral valve. Therefore, by extracting the outer contours of these two connections, the contours of the two leaflets can be extracted, and thus, according to the coordinates of each pixel point on the contours of the two extracted leaflets, the mitral valve opening distance corresponding to the current frame cardiac image can be obtained.

[0090] In an exemplary embodiment, the obtaining the position information of the mitral valve region of interest according to the current frame cardiac image includes:

[0091] Using an object detection model to detect the current frame cardiac image to obtain the position information of the mitral valve region of interest.

[0092] Thus, by using a pre-trained object detection model to detect the current frame cardiac image, the prediction result of the position where the mitral valve region of interest corresponding to the current frame cardiac image is located can be obtained. Specifically, the position information of the predicted mitral valve region of interest can be represented by the coordinates of the pixel point at the upper left corner and the lower right corner of the border of the predicted mitral valve region of interest.

[0093] Further, the using an object detection model to detect the current frame cardiac image to obtain the position information of the mitral valve region of interest includes:

[0094] Using an object detection model to detect the current frame cardiac image to obtain the position information of the candidate mitral valve region of interest;

[0095] According to the position information of the candidate mitral valve region of interest, calculating the position information of the mitral valve region of interest after magnifying by a preset multiple;

[0096] Taking the position information of the candidate mitral valve region of interest after magnifying by a preset multiple as the final position information of the mitral valve region of interest.

[0097] Although the use of the object detection model can detect the mitral valve region of interest in the current frame of the cardiac image and provide a preliminary positioning for the subsequent leaflet segmentation model, it will also cause the loss of detailed information such as the tissue around the mitral valve. Therefore, in the present invention, according to the position information of the candidate mitral valve region of interest, the position information of the candidate mitral valve region of interest after magnifying by a preset multiple is calculated, that is, the original bounding box of the candidate mitral valve region of interest detected by the object detection model is magnified by a preset multiple, such as 1.3 times, to obtain an enlarged bounding box, and the region defined by the enlarged bounding box is the final mitral valve region of interest. Since the region defined by the enlarged bounding box includes detailed information such as the tissue around the mitral valve, the segmentation accuracy of the subsequent leaflet segmentation model can be further improved. It should be noted that, as can be understood by those skilled in the art, the center position of the enlarged bounding box is the same as the center position of the original bounding box. Please refer to Figure 2 , which schematically shows a schematic diagram of the candidate mitral valve region of interest and the final mitral valve region of interest in the cardiac image provided by a specific example of the present invention. As Figure 2 shown, the region defined by the dashed bounding box in the figure is the candidate mitral valve region of interest; the region defined by the solid bounding box in the figure is the final mitral valve region of interest obtained by magnifying the candidate mitral valve region of interest (i.e., the dashed bounding box).

[0098] In an exemplary embodiment, after using the object detection model to detect the acquired current frame of the cardiac image to obtain the position information of the mitral valve region of interest, the method further includes:

[0099] Correcting the position information of the mitral valve region of interest according to the time sequence corresponding to the current frame of the cardiac image.

[0100] Thus, by correcting the position information of the mitral valve region of interest obtained by the object detection model, the extraction accuracy of the mitral valve region of interest can be further improved, and the accuracy of the subsequent leaflet segmentation can be further ensured. Specifically, the optical flow method can be used to correct the position information of the mitral valve region of interest to obtain the corrected position information of the mitral valve region of interest. It should be noted that, as can be understood by those skilled in the art, the optical flow method is a method that uses the change of pixels in the time domain in the image sequence and the correlation between adjacent frames to find the corresponding relationship between the previous frame and the current frame, so as to calculate the motion information of the object between adjacent frames.

[0101] In an exemplary embodiment, the segmenting the mitral valve region of interest corresponding to the current frame of the cardiac image by using the leaflet segmentation model according to the position information of the mitral valve region of interest to obtain the mitral valve leaflet mask image includes:

[0102] According to the position information of the mitral valve region of interest, crop the corresponding region on the current-frame cardiac image to obtain the mitral valve region of interest image;

[0103] Use the leaflet segmentation model to segment the mitral valve region of interest image to obtain the mitral valve leaflet mask image.

[0104] Thus, by first cropping the mitral valve region of interest on the current-frame cardiac image to obtain the mitral valve region of interest image, and then using the leaflet segmentation model to segment the mitral valve region of interest image, the computational load of the leaflet segmentation model can be further reduced, and thus the computational efficiency can be further improved. Please refer to Figure 3 and Figure 4 , wherein Figure 3 schematically shows a schematic diagram of the annotation of the mitral valve region of interest in the cardiac image provided by a specific example of the present invention, Figure 3 the region defined by the solid border in Figure 4 is the mitral valve region of interest; Figure 3 schematically shows the mitral valve leaflet mask image obtained by segmenting the mitral valve region of interest in the cardiac image shown in Figure 3 and Figure 4 . As shown in

[0105] By segmenting the mitral valve region of interest corresponding to the current-frame cardiac image, the mitral valve leaflet mask image can be accurately obtained, thus laying a good foundation for the subsequent calculation of the mitral valve opening distance.

[0106] Before using the leaflet segmentation model to segment the mitral valve region of interest image, the method further includes:

[0107] Adjust the size of the mitral valve region of interest image to a preset size.

[0108] Correspondingly, the using the leaflet segmentation model to segment the mitral valve region of interest image includes:

[0109] When the leaflet segmentation model is a neural network model, since the neural network model requires images of a unified size as input, by adjusting the size of the mitral valve region of interest image to a preset size, the input requirements of the leaflet segmentation model can be met. Specifically, the preset size can be set according to specific circumstances. As a preference, in the preset size, the dimension in the length direction of the image is consistent with the width direction, that is, the image after being adjusted to the preset size is a square image. For example, the preset size is 320*320. Thus, by setting the dimension in the length direction and the width direction in the preset size to be the same, it is more convenient to adjust the size of the mitral valve region of interest image to the preset size.

[0110] In an exemplary embodiment, the obtaining the mitral valve opening distance corresponding to the current-frame cardiac image according to the coordinates of each pixel point on the two leaflet contours includes:

[0111] Calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours:

[0112] Obtaining the mitral valve opening distance corresponding to the current-frame cardiac image according to the minimum pixel distance and the pre-obtained correspondence between the pixel distance and the physical distance.

[0113] Specifically, a pixel point at the upper left corner of the mitral valve leaflet mask image can be used as the origin, the width direction of the mitral valve leaflet mask image can be used as the X-axis (where the rightward direction is the positive direction of the X-axis), and the height direction of the mitral valve leaflet mask image can be used as the Y-axis (where the downward direction is the positive direction of the Y-axis) to create an image coordinate system, so as to obtain the coordinates of each pixel point on the two leaflet contours in the image coordinate system. Then, according to the coordinates of each pixel point on the two leaflet contours, the minimum pixel distance between the two leaflet contours can be calculated; and according to the calculated minimum pixel distance and the pre-obtained correspondence between the pixel distance and the physical distance, the minimum physical distance between the two leaflet contours can be obtained, that is, the mitral valve opening distance corresponding to the current-frame cardiac image.

[0114] Specifically, the border of the distance diameter line can be drawn according to the mitral valve opening distance diameter line outlined by the doctor in advance to obtain the coordinates of the pixel point at the upper left corner of the border and the length and width of the border, and calculate the pixel length of the diagonal of the border. The pixel length of the diagonal of the border is the pixel distance between the two leaflets. Then, according to the real distance (i.e., the physical distance) between the two leaflets measured by the doctor, the correspondence between the pixel distance and the physical distance can be obtained. Please refer to Figure 5 and Figure 6 , whereFigure 5 Schematically shown is a schematic diagram of a doctor outlining the mitral valve opening spacing diameter provided by a specific example of the present invention; Figure 6 Shown is a schematic diagram of the drawing of the border of the spacing diameter provided by a specific example of the present invention. As Figure 5 and Figure 6 shown, Figure 5 and Figure 6 the white solid lines in Figure 6 represent the mitral valve opening spacing diameter, and the dashed box in

[0115] Please refer to Figure 7 , which schematically shows a specific flow diagram of calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention. As Figure 7 shown, in this embodiment, calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours includes:

[0116] Step A: Determine a pixel point on each of the two leaflet contours as a starting point according to the coordinates of each pixel point on the two leaflet contours;

[0117] Step B: Draw a circle with the two determined starting points as the two endpoints of a diameter;

[0118] Step C: Determine whether the drawn circle has new intersections with the two leaflet contours. If not, execute Step D; if so, execute Step E;

[0119] Step D: Take the diameter of the drawn circle as the minimum pixel distance between the two leaflet contours;

[0120] Step E: Determine whether the number of the new intersections is greater than or equal to 2. If so, execute Step E1; if not, execute Step E2;

[0121] Step E1: Draw a circle with two of the new intersections on different leaflet contours as the two endpoints of a new diameter, and return to execute Step C;

[0122] Step E2: Draw a circle with the new intersection and the original intersection on the other leaflet contour as the two endpoints of a new diameter, and return to execute Step C.

[0123] This embodiment mainly uses the circle tangent point geometric method to calculate the mitral valve opening spacing. Specifically, please refer to Figure 8 , which schematically shows the principle diagram of calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention. As Figure 8As shown, the starting point determined on the upper leaflet contour is point A, and the starting point determined on the lower leaflet contour is point B. A circle with points A and B as the two endpoints of the diameter intersects the upper leaflet contour at a new intersection point C and the lower leaflet contour at a new intersection point D. Since the circle with points A and B as the two endpoints of the diameter intersects the upper and lower leaflet contours at a total of 2 new intersection points, then it is necessary to continue drawing a circle with these two new intersection points C and D as the two endpoints of a new diameter. If the circle drawn with these two new intersection points C and D as the two endpoints of the new diameter does not intersect the upper and lower leaflet contours (i.e., the circle drawn with these two new intersection points C and D as the two endpoints of the new diameter is tangent to the upper and lower leaflet contours), then the distance between points C and D is used as the minimum pixel distance between the upper and lower leaflet contours. If the circle drawn with these two new intersection points C and D as the two endpoints of the new diameter intersects the upper leaflet contour at at least one new intersection point and the lower leaflet contour at at least one new intersection point, then a circle is continued to be drawn with one of the new intersection points on the upper leaflet contour and one of the new intersection points on the lower leaflet contour as the two endpoints of the new diameter; if the circle drawn with these two new intersection points C and D as the two endpoints of the new diameter intersects only the upper leaflet contour at one new intersection point, then a circle is continued to be drawn with this new intersection point and point D (the original intersection point) as the two endpoints of the new diameter; if the circle drawn with these two new intersection points C and D as the two endpoints of the new diameter intersects only the lower leaflet contour at one new intersection point, then a circle is continued to be drawn with this new intersection point and point C (the original intersection point) as the two endpoints of the new diameter; if the drawn circle still intersects the upper and lower leaflet contours at new intersection points, then circles are continued to be drawn based on the new intersection points until the drawn circle is tangent to both the upper and lower leaflet contours. Please continue to refer to Figure 9 , which schematically shows a result diagram of calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention. The pixel length of the white straight line in the figure is the minimum pixel distance between the upper and lower leaflet contours. As Figure 9As shown, when the constructed circle has no new intersection points (i.e., is tangent) with both the upper leaflet contour and the lower leaflet contour, stop constructing the circle, and use the diameter of this circle as the minimum pixel distance between the upper leaflet contour and the lower leaflet contour. It should be noted that, as can be understood by those skilled in the art, after constructing a circle with a certain diameter, the coordinates of each pixel point on the contour of the constructed circle can be obtained. The coordinates of each pixel point on the circle are represented by set M, and the coordinates of each pixel point on the two leaflet contours are represented by set N. The union of set M and set N is the coordinate set of the intersection points of the circle and the two leaflet contours. That is, all the intersection points (including the two endpoints on the diameter) of the circle and the two leaflet contours can be obtained according to the union of set M and set N. If the number of intersection points corresponding to the union of set M and set N is greater than 2, continue the next iteration; if the number of intersection points corresponding to the union of set M and set N is equal to 2, stop the iteration, and the diameter of the constructed circle is the minimum pixel distance between the upper leaflet contour and the lower leaflet contour. Please continue to refer to Figure 10 , which schematically shows a schematic diagram of the iterative process for calculating the minimum pixel distance between two leaflet contours provided by the first embodiment of the present invention. As Figure 10 shown, the gray circle contour in the figure is the circle constructed during continuous iteration, and the white solid line is the diameter line formed by the new intersection points of the circle with a smaller diameter in the figure and the upper leaflet contour and the new intersection points with the lower leaflet contour, that is, the diameter of the next iterative circle.

[0124] Furthermore, based on the coordinates of each pixel point on the two leaflet contours, one pixel point is respectively determined as the starting point on each of the two leaflet contours, including:

[0125] For each leaflet contour, according to the coordinates of each pixel point on this leaflet contour, the pixel point located on the leftmost side is used as the starting point on this leaflet contour; or

[0126] For each leaflet contour, according to the coordinates of each pixel point on this leaflet contour, the pixel point with the smallest sum of the X coordinate and the Y coordinate is used as the starting point on this leaflet contour.

[0127] Thus, for each leaflet contour, by using the pixel point located on the leftmost side of this leaflet contour or the pixel point with the smallest sum of the X coordinate and the Y coordinate on this leaflet contour as the starting point, the number of iterations can be effectively reduced, so that the minimum pixel distance between the two leaflet contours can be calculated more quickly, and the calculation efficiency of the mitral valve opening distance can be improved. Specifically, for each leaflet contour, sort the X coordinates of all the pixel points on this leaflet contour, and the pixel point with the smallest X coordinate is the pixel point located on the leftmost side of this leaflet contour.

[0128] Even further, for step E1, the method further includes:

[0129] Determine whether the absolute value of the difference between the distance between the two new intersection points located on different leaflet contours and the diameter of the currently drawn circle is less than a second preset threshold;

[0130] If so, take the distance between the two new intersection points as the minimum pixel distance between the two leaflet contours;

[0131] If not, draw a circle with the two new intersection points as the two endpoints of a new diameter;

[0132] For step E2, the method further includes:

[0133] Determine whether the absolute value of the difference between the distance between the new intersection point and the original intersection point located on another leaflet contour and the diameter of the currently drawn circle is less than the second preset threshold;

[0134] If so, take the distance between the new intersection point and the original intersection point located on another leaflet contour as the minimum pixel distance between the two leaflet contours;

[0135] If not, draw a circle with the new intersection point and the original intersection point located on another leaflet contour as the two endpoints of a new diameter.

[0136] Thus, for the case where the number of new intersection points is greater than or equal to 2, only when the absolute value of the difference between the distance between the two new intersection points located on different leaflet contours and the diameter of the currently drawn circle is greater than or equal to the second preset threshold, continue to draw a circle with these two new intersection points as the two endpoints of a new diameter; for the case where the number of new intersection points is equal to 1, only when the absolute value of the difference between the distance between the new intersection point and the original intersection point located on another leaflet contour and the diameter of the currently drawn circle is greater than or equal to the second preset threshold, continue to draw a circle with this new intersection point and the original intersection point located on another leaflet contour as the two endpoints of a new diameter, thereby effectively reducing the number of iterations and further improving the detection efficiency of the mitral valve opening distance.

[0137] Please continue to refer to Figure 11 , which schematically shows a specific flowchart of calculating the minimum pixel distance between two leaflet contours provided by the second embodiment of the present invention. As Figure 11 shown, calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours includes:

[0138] According to the coordinates of each pixel point on the two leaflet contours, a decision boundary for distinguishing the two leaflet contours and support vectors located on each of the leaflet contours are determined using a support vector machine, where the support vector is the pixel point on the leaflet contour that is closest to the decision boundary;

[0139] For each of the support vectors, a perpendicular line is drawn from the support vector to the decision boundary. If the perpendicular line intersects with the other leaflet contour, then this support vector is used as a target support vector;

[0140] For each of the target support vectors, the pixel distance between the target support vector and the corresponding intersection point is calculated;

[0141] The smallest calculated pixel distance is used as the minimum pixel distance between the two leaflet contours.

[0142] Specifically, please refer to Figure 12 , which schematically shows a schematic diagram of the principle for calculating the minimum pixel distance between two leaflet contours provided by the second embodiment of the present invention. As Figure 12As shown, by using a support vector machine, binary classification can be achieved for each pixel point on the leaflet contour (i.e., distinguishing which pixel points are on the upper leaflet contour and which are on the lower leaflet contour). The solid gray line is the determined decision boundary, and the two dashed lines are the upper and lower boundaries of the two leaflet contours. Pixel points above the upper boundary are pixel points on the upper leaflet contour, and pixel points below the lower boundary are pixel points on the lower leaflet contour. Pixel points on the upper boundary (i.e., the pixel points on the upper leaflet contour closest to the decision boundary) are the support vectors on the upper leaflet contour, and pixel points on the lower boundary (i.e., the pixel points on the lower leaflet contour closest to the decision boundary) are the support vectors on the lower leaflet contour. After determining the decision boundary and support vectors, for each support vector on the upper leaflet contour, draw a perpendicular line from this support vector to the decision boundary. If the drawn perpendicular line intersects the lower leaflet contour, then take this support vector as the target support vector and calculate the pixel distance between this target support vector and the intersection point (if there are multiple intersection points, calculate the pixel distance between this target support vector and each intersection point respectively, or only calculate the pixel distance between the intersection point closest to this target support vector and this target support vector); similarly, for each support vector on the lower leaflet contour, draw a perpendicular line from this support vector to the decision boundary. If the drawn perpendicular line intersects the upper leaflet contour, then take this support vector as the target support vector and calculate the pixel distance between this target support vector and the intersection point (if there are multiple intersection points, calculate the pixel distance between this target support vector and each intersection point respectively, or only calculate the pixel distance between the intersection point closest to this target support vector and this target support vector). Finally, sort the calculated pixel distances, and the smallest calculated pixel distance is the minimum pixel distance between the upper leaflet contour and the lower leaflet contour.

[0143] Further, determining the decision vectors on each of the leaflet contours and the support vectors on each of the leaflet contours includes:

[0144] Perform binary classification on each pixel point on the two leaflet contours to obtain the pixel points on the upper leaflet contour and the pixel points on the lower leaflet contour, so as to determine the decision boundary;

[0145] Take the pixel points on the upper leaflet contour closest to the decision boundary as the support vectors on the upper leaflet contour, and take the pixel points on the lower leaflet contour closest to the decision boundary as the support vectors on the lower leaflet contour.

[0146] Specifically, a Support Vector Machine (SVM) can be used to perform binary classification on the coordinates of each point on the leaflet contour. Since there is a decision boundary in the feature space where each pixel point is located, that is, the classification boundary that divides the pixel points according to the upper leaflet contour and the lower leaflet contour. As Figure 12 shown, all pixel points above the decision boundary (classification boundary) belong to the upper leaflet contour, and the coordinate points below the boundary belong to the lower leaflet contour. This classification boundary (decision boundary) needs to maximize the margins on each side to reduce the classification error rate.

[0147] Among them, the formula of the decision boundary can be expressed as:

[0148] W T X + b = 0

[0149] In the formula, W is the feature weight vector, which is the normal vector of the decision boundary, b is the bias value, which is the intercept of the decision boundary, and X is the set composed of the coordinates of each pixel point. The decision boundary that satisfies the margin maximization constitutes two parallel upper and lower boundaries. The formula of the upper boundary can be expressed as:

[0150] W T X + b = 1

[0151] The formula of the lower boundary can be expressed as:

[0152] W T x + b = -1

[0153] Among them, the pixel points located on the upper boundary are the support vectors on the upper leaflet contour, and the pixel points located on the lower boundary are the support vectors on the lower leaflet contour. After determining the support vectors on the upper leaflet contour and the support vectors on the lower leaflet contour, for each support vector on the upper leaflet contour, draw a perpendicular line from this support vector to the decision boundary. If the drawn perpendicular line intersects with the lower leaflet contour, then regard this support vector as the target support vector, and calculate the pixel distance between this target support vector and the intersection point; similarly, for each support vector on the lower leaflet contour, draw a perpendicular line from this support vector to the decision boundary. If the drawn perpendicular line intersects with the upper leaflet contour, then regard this support vector as the target support vector, and calculate the pixel distance between this target support vector and the intersection point. Finally, sort the calculated pixel distances, and the smallest calculated pixel distance is the minimum pixel distance between the upper leaflet contour and the lower leaflet contour.

[0154] In an exemplary embodiment, the method further includes:

[0155] According to the time sequence corresponding to each frame of cardiac image and the mitral valve opening distance, an opening distance spectrogram is drawn to represent the corresponding relationship between the number of frames and the mitral valve opening distance.

[0156] Please refer to Figure 13 , which schematically shows the opening distance spectrogram provided by a specific example of the present invention. As Figure 13 shown, the horizontal axis of the opening distance spectrogram is the number of frames, and the vertical axis is the mitral valve opening distance. Thus, by drawing the opening distance spectrogram, the mitral valve opening distance corresponding to each frame of cardiac image can be displayed, making it more convenient for doctors to view the detection results.

[0157] In an exemplary embodiment, after obtaining the mitral valve opening distances corresponding to all frames of cardiac images, the method further includes:

[0158] Drawing two valve leaf contours, the mitral valve opening distance diameter line, and the text content of the maximum mitral valve opening distance on the cardiac image corresponding to the maximum mitral valve opening distance and outputting.

[0159] Please continue to refer to Figure 14 , which schematically shows the drawing schematic diagram of the cardiac image corresponding to the maximum mitral valve opening distance provided by a specific example of the present invention. As Figure 14 shown, the area defined by the rectangular frame in the figure is the mitral valve region of interest. The two white curves in the figure represent the two valve leaf contours. The white straight line in the figure is the mitral valve opening distance diameter line. The text "dist = 2.82" (unit: cm) in the upper left of the figure is the maximum mitral valve opening distance. Thus, by drawing two valve leaf contours, the mitral valve opening distance diameter line, and the text content of the maximum mitral valve opening distance on the cardiac image corresponding to the maximum mitral valve opening distance, it is convenient for doctors to more intuitively view the detection results, enabling doctors to make a diagnosis on whether the patient's mitral valve is stenotic based on the maximum mitral valve opening distance, which is more conducive to improving the accuracy of doctors' diagnosis.

[0160] In an exemplary embodiment, after drawing two valve leaf contours, the mitral valve opening distance diameter line, and the text content of the maximum mitral valve opening distance on the cardiac image corresponding to the maximum mitral valve opening distance, the method further includes:

[0161] Performing denoising processing on the cardiac image to filter out the noise on the cardiac image.

[0162] Specifically, the median filtering method can be adopted to set the gray value of each pixel point in the cardiac image to the median of the gray values of all pixel points within the neighborhood window of this pixel point. The size parameter of the filtering kernel can be set according to specific circumstances, for example, set to 5×5. Thus, by adopting the median filtering method, the salt-and-pepper noise in the cardiac image can be effectively removed. It should be noted that as can be understood by those skilled in the art, in some other embodiments, other filtering methods other than the median filtering method can also be adopted to perform filtering processing on the cardiac image, and the present invention does not limit this.

[0163] In an exemplary embodiment, the target detection model is a ResNet50 neural network model. Since skip connections (or shortcuts) are used in ResNet, it directly passes the activation value of a certain network layer to a deeper layer of the network. In addition, the skip connection only passes data. Through the skip connection, the signal can be transmitted without attenuation during backpropagation, without worrying about the change of the gradient, and can transmit effective gradients to the upper layer. Thus, through the skip connection, the problem of gradient disappearance caused by deepening the network layer can be effectively alleviated. By stacking Residual blocks, a very deep network model can be constructed, enabling effective training of deep network layers.

[0164] Furthermore, the target detection model includes a first convolutional layer, a first pooling layer, a plurality of cascaded residual sub-networks, a second pooling layer, and a fully connected layer. Among them, the first convolutional layer is used to extract mitral valve features from the input current-frame cardiac image, the first pooling layer is used to perform a pooling operation on the output of the first convolutional layer, the residual sub-network is used to extract mitral valve features from the output of the first pooling layer or the output of the previous-level residual sub-network, the second pooling layer is used to perform a pooling operation on the output of the last-level residual sub-network, and the fully connected layer is used to perform non-linear mapping regression on the output of the second pooling layer to obtain the position information of the mitral valve region of interest.

[0165] Even further, each of the residual sub-networks includes a plurality of cascaded residual modules, and each of the residual modules includes a plurality of cascaded second convolutional layers. Among them, the input of the first-level second convolutional layer is added to the output of the last-level second convolutional layer and used as the output of the residual module.

[0166] Further, the size of the convolutional kernel of the second convolutional layer at the first level and the size of the convolutional kernel of the second convolutional layer at the last level are both 1×1.

[0167] Specifically, please refer to Figure 15 , which schematically shows the structural diagram of the target detection model provided by a specific example of the present invention. AsFigure 15 As shown, in this example, the target detection model includes a first convolutional layer, a first pooling layer, a first residual sub-network, a second residual sub-network, a third residual sub-network, a fourth residual sub-network, a second pooling layer, and a fully connected layer. Among them, the first convolutional layer is used to extract mitral valve features from the input current-frame cardiac image, the first pooling layer is used to perform a pooling operation on the output of the first convolutional layer, the first residual sub-network is used to extract mitral valve features from the output of the first pooling layer, the second residual sub-network is used to extract mitral valve features from the output of the first residual sub-network, the third residual sub-network is used to extract mitral valve features from the output of the second residual sub-network, the fourth residual sub-network is used to extract mitral valve features from the output of the third residual sub-network, the second pooling layer is used to perform a pooling operation on the output of the fourth residual sub-network, and the fully connected layer is used to perform non-linear mapping regression on the output of the second pooling layer to obtain the position information of the mitral valve region of interest. Further, the first pooling layer is a max pooling layer, and the second pooling layer is an average pooling layer.

[0168] Further, the first residual sub-network includes 3 cascaded first residual modules, the second residual sub-network includes 4 cascaded second residual modules, the third residual sub-network includes 6 cascaded residual modules C1, and the fourth residual sub-network includes 3 cascaded fourth residual modules. Please continue to refer to Figure 16a , which schematically shows the structural diagram of the first residual module provided by a specific example of the present invention. As Figure 16a shown, the first residual module includes 3 cascaded second convolutional layers, namely second convolutional layer A1, second convolutional layer A2, and second convolutional layer A3. Among them, the size of the convolutional kernel of the second convolutional layer A1 is 1×1, the number of output channels is 64, the stride is 1, the size of the convolutional kernel of the second convolutional layer A2 is 3×3, the number of output channels is 64, the stride is 1, the size of the convolutional kernel of the second convolutional layer A3 is 1×1, the number of output channels is 256, the stride is 1, and the identity mapping of the input of the second convolutional layer A1 is added to the output of the second convolutional layer A3 as the output of the first residual module. Please continue to refer to Figure 16b , which schematically shows the structural diagram of the second residual module provided by a specific example of the present invention. As Figure 16bAs shown, the second residual module includes three cascaded second convolutional layers, namely the second convolutional layer B1, the second convolutional layer B2, and the second convolutional layer B3. Among them, the size of the convolutional kernel of the second convolutional layer B1 is 1×1, the number of output channels is 128, the stride is 1, the size of the convolutional kernel of the second convolutional layer B2 is 3×3, the number of output channels is 128, the stride is 2, the size of the convolutional kernel of the second convolutional layer B3 is 1×1, the number of output channels is 512, the stride is 1. The identity mapping of the input of the second convolutional layer B1 is added to the output of the second convolutional layer B3 and then used as the output of the second residual module. Please continue to refer to Figure 16c , which schematically shows the structural diagram of the third residual module provided by a specific example of the present invention. As shown in Figure 16c As shown, the third residual module includes three cascaded second convolutional layers, namely the second convolutional layer C1, the second convolutional layer C2, and the second convolutional layer C3. Among them, the size of the convolutional kernel of the second convolutional layer C1 is 1×1, the number of output channels is 256, the stride is 1, the size of the convolutional kernel of the second convolutional layer C2 is 3×3, the number of output channels is 256, the stride is 2, the size of the convolutional kernel of the second convolutional layer C3 is 1×1, the number of output channels is 1024, the stride is 1. The identity mapping of the input of the second convolutional layer C1 is added to the output of the second convolutional layer C3 and then used as the output of the third residual module. Please continue to refer to Figure 16d , which schematically shows the structural diagram of the fourth residual module provided by a specific example of the present invention. As shown in Figure 16d As shown, the fourth residual module includes three cascaded second convolutional layers, namely the second convolutional layer D1, the second convolutional layer D2, and the second convolutional layer D3. Among them, the size of the convolutional kernel of the second convolutional layer D1 is 1×1, the number of output channels is 512, the stride is 1, the size of the convolutional kernel of the second convolutional layer D2 is 3×3, the number of output channels is 512, the stride is 2, the size of the convolutional kernel of the second convolutional layer D3 is 1×1, the number of output channels is 2048, the stride is 1. The identity mapping of the input of the second convolutional layer D1 is added to the output of the second convolutional layer D3 and then used as the output of the fourth residual module.

[0169] Since the skip connections between the second convolutional layer A1 and the second convolutional layer A3, the second convolutional layer B1 and the second convolutional layer B3, the second convolutional layer C1 and the second convolutional layer C3, and the second convolutional layer D1 and the second convolutional layer D3 all adopt identity mapping connections, the training speed of the target detection model can be accelerated and the training effect of the target detection model can be improved without increasing additional parameters and computational complexity.

[0170] Further, the samples used in the training process of the object detection model are sample cardiac images with the mitral valve region of interest (ROI) already marked. Specifically, based on the sample cardiac images in which doctors have previously outlined the ROI, the threshold segmentation method can be used to segment the ROI in the sample cardiac images to extract the contour features of all levels (including the outer contour and inner contour of the mitral valve leaflets), and obtain the coordinates of the upper left pixel point and the lower right pixel point of the border of each contour (the X coordinate of the upper left pixel point plus the width of the border is the X coordinate of the lower right pixel point, and the Y coordinate of the upper left pixel point plus the length of the border is the Y coordinate of the lower right pixel point). The coordinates of the upper left pixel point with the smallest total coordinate value (the sum of the X coordinate and the Y coordinate) are used as the coordinates of the upper left pixel point of the bounding box of the mitral valve ROI, and the coordinates of the lower right pixel point with the largest total coordinate value (the sum of the X coordinate and the Y coordinate) are used as the coordinates of the lower right pixel point of the bounding box of the mitral valve ROI. Thus, based on the coordinates of the upper left and lower right pixel points of the bounding box of the mitral valve ROI, the mitral valve ROI can be marked in the sample cardiac images outlined by doctors in advance to obtain the samples required for the training of the object detection model. Please refer to Figure 17 , which schematically shows the labeled sample cardiac image provided by a specific example of the present invention. As Figure 17 shown, the area defined by the solid line box in the figure is the mitral valve ROI outlined by doctors in advance, and the area defined by the dotted line box in the figure is the finally marked mitral valve ROI. The position information of the finally marked mitral valve ROI is represented by the coordinates of the upper left pixel point A and the lower right pixel point B of the dotted line box.

[0171] It should be noted that, as can be understood by those skilled in the art, since the object detection model requires images of a unified size as input, it is necessary to convert the sample cardiac images with the mitral valve ROI already marked to a preset size, such as 320×320.

[0172] In an exemplary embodiment, the loss function used in the training process of the object detection model is Focal Loss, and the formula of Focal Loss is as follows:

[0173] FL(P t )=-(1-P t ) γ log(P t )

[0174] In the formula, P t represents the predicted probability, (1-P t ) γ is an adjustable factor, and γ is an adjustable focusing parameter.

[0175] Thus, by introducing the Focal Loss function with (1 - P t ) γ as an adjustable factor as the loss function during the model training process, the degree of reduction of the weights of easily classifiable samples can be adjusted, which is more convenient for the training of the model.

[0176] The model parameters of the object detection model include two categories: feature parameters and hyperparameters. The feature parameters are continuously iteratively learned by the neural network model and are used to learn image features, such as mitral valve features. The feature parameters include weight parameters and bias parameters. The hyperparameters are parameters set manually during training. Only by setting appropriate hyperparameters can the feature parameters be learned from the samples. The hyperparameters can include the learning rate, the number of hidden layers, the size of the convolution kernel, the number of training iterations, and the batch size for each iteration. During specific training, the training samples are loaded into the pre-constructed neural network model, then the parameters in the network model are initially set, then the network is initialized, and finally the network model is run for training. After training for a certain time, it is judged whether the loss function converges. If it does not converge, continue training until the loss function converges, then the training process is completed, and the corresponding weight parameters at this time are saved. During the training process, the stochastic gradient descent method can be used to update the weight parameters of the network. As an example, the present invention sets the learning rate to 1e-5 (i.e., 0.00001) and uses a callback function to update the learning rate. When it is found that the loss function of the validation set no longer decreases after 2 epochs (training cycles), the learning rate is divided by 10. At the same time, EarlyStopping (early stopping method) is set to intercept and save the parameter model with the best results after monitoring the mean average precision (mAP) for 15 epochs (training cycles) to prevent overfitting. Among them, the calculation formula for the mean average precision (mAP) is as follows:

[0177]

[0178] In the formula, P is the accuracy rate, and R is the recall rate.

[0179] Furthermore, the present invention also uses a test set to test the trained object detection model to evaluate the algorithm accuracy of the object detection model. Specifically, the intersection over union (IOU) of the predicted bounding box (the bounding box of the predicted mitral valve region of interest) and the ground truth bounding box (the bounding box of the actual mitral valve region of interest) of the object detection model can be calculated to evaluate the algorithm accuracy of the object detection model. Among them, the calculation formula for the intersection over union (IOU) is as follows:

[0180]

[0181] Wherein, A is the predicted bounding box output after the test samples in the test set are detected by the target detection model, and B is the marked actual bounding box (i.e., the ground truth bounding box).

[0182] In an exemplary embodiment, the leaflet segmentation model is a DenseNet neural network model. Since the DenseNet neural network model is a convolutional neural network with dense connections, the input of each layer comes from the outputs of all previous layers. This neural network structure strengthens the transmission of features and makes more effective use of features. In addition, the DenseNet neural network model has good anti-overfitting performance and is especially suitable for applications with relatively scarce training data. Therefore, using the DenseNet neural network model as the leaflet segmentation model in the present invention can effectively improve the segmentation efficiency and accuracy of the mitral valve. Specifically, the DenseNet neural network model is composed of multiple dense connection blocks connected by transition blocks, that is, any two adjacent dense connection blocks are connected by a transition block, and the number of convolutional output channels in the dense connection block is the same, so as to be able to superimpose the feature information of each layer.

[0183] One layer in the dense connection block is called a bottleneck layer. The dense connections in DenseNet connect each layer in a dense connection block to all subsequent layers to achieve feature reuse. Please refer to Figure 18 , which schematically shows the structural diagram of the dense connection block provided by a specific example of the present invention. As Figure 18 shown, assume that a dense connection block has L bottleneck layers, X0 is the input of the dense connection block (defined as the output of the 0th layer), and the lth layer takes the outputs X0, ……, X L-1 of all previous layers as inputs, that is, the following relationship is satisfied between the input of the lth layer and the outputs of all previous layers:

[0184] X l = H L ([X0, X1, … X L-1 )

[0185] where, [X0, X1, … X L-1 represents the combination connection of the outputs from the 0th layer to the (L - 1)th layer as the input of the Lth bottleneck layer, and H L represents all the operations of the Lth bottleneck layer. Specifically, each of the bottleneck layers includes multiple operations: batch normalization BN, ReLU activation function, and 3×3 convolution.

[0186] Please continue to refer to Figure 19 , which schematically shows the structural diagram of the leaflet segmentation model provided by a specific example of the present invention. As Figure 19As shown, in this example, the leaflet segmentation model includes a third convolutional layer, a third pooling layer, a first dense connection block, a first transition block, a second dense connection block, a second transition block, a third dense connection block, a third transition block, a fourth dense connection block, a first upsampling transition block, a second upsampling transition block, and a fourth convolutional layer that are connected in sequence. The third pooling layer is preferably a max pooling layer. Among them, the third convolutional layer is used to extract mitral valve features from the input image (i.e., the image of the mitral valve region of interest corresponding to the current frame of cardiac image). The third pooling layer is used to perform a pooling operation on the output of the third convolutional layer to remove unnecessary redundant information in the image. The first dense connection block is used to extract mitral valve features from the output of the third pooling layer. The first transition block is used to perform a compression operation on the output of the first dense connection block to reduce the size of the feature map output by the first dense connection block. The second dense connection block is used to extract mitral valve features from the output of the first transition block. The second transition block is used to perform a compression operation on the output of the second dense connection block to reduce the size of the feature map output by the second dense connection block. The third dense connection block is used to extract mitral valve features from the output of the second transition block. The third transition block is used to perform a compression operation on the output of the third dense connection block to reduce the size of the feature map output by the third dense connection block. The first upsampling transition block is used to perform a deconvolution operation on the output of the fourth dense connection block to increase the size of the feature map output by the fourth dense connection block. The second upsampling transition block is used to perform a deconvolution operation on the output of the first upsampling transition block to increase the size of the feature map output by the first upsampling transition block. The fourth convolutional layer is used to perform a non-linear mapping regression on the output of the second upsampling transition block to obtain the segmentation result of the mitral valve leaflets.

[0187] Specifically, the fourth convolutional layer can perform a non-linear mapping regression on the output of the second upsampling transition block through the sigmoid function. The formula of the sigmoid function is as follows:

[0188]

[0189] As can be seen from the above formula, the sigmoid function can map any input real number to the real number mapping interval (0, 1). When the input value x is large, the output value g tends to 1, and when the input value x is small, the output value g tends to 0.

[0190] It should be noted that, as can be understood by those skilled in the art, the first dense connection block, the second dense connection block, the third dense connection block, and the fourth dense connection block all include a plurality of bottleneck layers, and the number of bottleneck layers in the first dense connection block, the second dense connection block, the third dense connection block, and the fourth dense connection block may be the same or different. The specific number can be set according to actual needs, and the present invention does not limit this. For example, 6 bottleneck layers can be provided in the first dense connection block, 12 bottleneck layers can be provided in the second dense connection block, 24 bottleneck layers can be provided in the third dense connection block, and 16 bottleneck layers can be provided in the fourth dense connection block.

[0191] Please continue to refer to Figure 20 , which schematically shows the structural diagram of the bottleneck layer provided by a specific example of the present invention. As Figure 20 shown, the bottleneck layer includes a first batch normalization layer A, a first activation layer A, a fifth convolutional layer A, a first batch normalization layer B, a first activation layer B, and a fifth convolutional layer B connected in sequence. Among them, the size of the convolutional kernel of the fifth convolutional layer A is 1×1, and the size of the convolutional kernel of the fifth convolutional layer B is 3×3. Thus, the present invention can reduce the number of feature maps and the dimension of each feature map by adding a 1×1 convolution before the 3×3 convolution in the bottleneck layer to reduce the computational amount, and can also fuse the features of each channel. In addition, since the bottleneck layer performs batch normalization BN operation and ReLU activation operation before performing 1×1 and 3×3 convolution operations, the training speed and convergence efficiency can be further improved.

[0192] Please continue to refer to Figure 21 , which schematically shows the structural diagram of the transition block provided by a specific example of the present invention. As Figure 21 shown, the first transition block, the second transition block, and the third transition block all include a second batch normalization layer, a second activation layer, a sixth convolutional layer, and a fourth pooling layer connected in sequence. The fourth pooling layer is preferably an average pooling layer. Among them, the size of the convolutional kernel of the fifth convolutional layer is 1×1. Thus, through the convolution operation of the fifth convolutional layer, the dimension of the feature map can be reduced, and through the average pooling operation of the fourth pooling layer, the problem of too many channels of the feature map can be solved, and the problem of model complexity caused by too many dense connection blocks can be prevented. In addition, since each transition block also performs batch normalization BN operation and ReLU activation operation before performing 1×1 convolution operation, the number of parameters can be further compressed.

[0193] Please continue to refer to Figure 22 , which schematically shows the structural diagram of the upward transition block provided by a specific example of the present invention. As Figure 22As shown, both the first upward transition block and the second upward transition block include a third batch normalization layer A, a third activation layer A, a seventh convolution layer A, a third batch normalization layer B, a third activation layer B, a seventh convolution layer B, a third batch normalization layer C, a third activation layer C, and a deconvolution layer that are connected in sequence. Among them, the sizes of the convolution kernels of the seventh convolution layer A and the seventh convolution layer B are both 3×3.

[0194] Further, the samples used in the training process of the leaflet segmentation model are mitral valve region of interest sample images and mitral valve leaflet mask images corresponding to the mitral valve region of interest sample images. Specifically, the OpenCV contour extraction algorithm can be used to find the mitral valve contour in the mitral valve region of interest sample image to segment the mitral valve leaflet mask image.

[0195] It should be noted that as can be understood by those skilled in the art, since the leaflet segmentation model requires images of the same size as input, it is necessary to convert both the mitral valve region of interest sample image and the mitral valve leaflet mask image corresponding to the mitral valve region of interest sample image to a preset size, such as 320×320.

[0196] In an exemplary embodiment, the loss function used in the training process of the leaflet segmentation model is the binary_crossentropy cross-entropy loss function, and the formula of the binary_crossentropy cross-entropy loss function is as follows:

[0197]

[0198]

[0199] In the formula, y i is the true label, is the prediction result.

[0200] Further, after the training of the leaflet segmentation model is completed, the present invention also uses the Dice coefficient formula to evaluate the algorithm accuracy of the leaflet segmentation model. The Dice coefficient formula is as follows:

[0201]

[0202] In the formula, X is the prediction result and Y is the true label.

[0203] As an example, during the training process of the leaflet segmentation model, the learning rate is set to 1e-3 (i.e., 0.001), and Adam (adaptive moment estimation) is used as the optimizer. The learning rate of each parameter is dynamically adjusted using the first-order moment estimation and second-order moment estimation of the gradient, and clipnorm = 0.001 is added to the parameters of the optimizer to clip the gradient.

[0204] Based on the same inventive concept, the present invention also provides an electronic device. Please refer to Figure 23 , which schematically shows the block structure diagram of the electronic device provided by an embodiment of the present invention. As Figure 23 shown, the electronic device includes a processor 101 and a memory 103. A computer program is stored on the memory 103. When the computer program is executed by the processor 101, the mitral valve opening distance detection method described above is implemented. Since the electronic device provided by the present invention and the mitral valve opening distance detection method described above belong to the same inventive concept, the electronic device provided by the present invention has all the advantages of the mitral valve opening distance detection method described above. Therefore, the advantages of the electronic device provided by the present invention will not be described one by one.

[0205] As Figure 23 shown, the electronic device further includes a communication interface 102 and a communication bus 104. Among them, the processor 101, the communication interface 102, and the memory 103 complete mutual communication through the communication bus 104. The communication bus 104 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus 104 can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface 102 is used for communication between the above-mentioned electronic device and other devices.

[0206] The processor 101 mentioned in the present invention may be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 101 is the control center of the electronic device, connecting all parts of the electronic device through various interfaces and circuits.

[0207] The memory 103 can be used to store the computer program. The processor 101 realizes various functions of the electronic device by running or executing the computer program stored in the memory 103 and calling the data stored in the memory 103.

[0208] The memory 103 may include non-volatile and / or volatile memory. The non-volatile memory may include Read-Only Memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. The volatile memory may include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Rambus Dynamic RAM (RDRAM), etc.

[0209] The present invention also provides a readable storage medium with a computer program stored therein. When the computer program is executed by a processor, it can implement the mitral valve opening distance detection method described above. Since the readable storage medium provided by the present invention and the mitral valve opening distance detection method described above belong to the same inventive concept, the readable storage medium provided by the present invention has all the advantages of the mitral valve opening distance detection method described above. Therefore, the advantages of the readable storage medium provided by the present invention will not be elaborated one by one.

[0210] The readable storage medium according to the embodiments of the present invention may adopt any combination of one or more computer-readable media. The readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer hard disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device.

[0211] The computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which computer-readable program code is carried. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination of the above.

[0212] The computer program code for performing the operations of the present invention can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user computer, partially on the user computer, executed as an independent software package, partially on the user computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0213] In summary, compared with the prior art, the mitral valve opening distance detection method, electronic device, and storage medium provided by the present invention have the following advantages:

[0214] In the present invention, first, according to the acquired current-frame cardiac motion image, the position information of the mitral valve region of interest is obtained; then, according to the position information of the mitral valve region of interest, a leaflet segmentation model is used to segment the mitral valve region of interest corresponding to the current-frame cardiac motion image to obtain a mitral valve leaflet mask image; finally, connected component analysis is performed on the mitral valve leaflet mask image, and according to the analysis result of the connected components, the mitral valve opening distance corresponding to the current-frame cardiac motion image is obtained. Thus, it can be seen that the present invention can automatically detect the mitral valve opening distance, provide an evaluation basis for the diagnosis of whether the mitral valve is stenotic, not only improve the overall algorithm accuracy, but also reduce the differential problems that may be caused by human factors, and further can better assist doctors in improving the diagnosis efficiency and effectively reduce the risk of misdiagnosis in the process of analyzing mitral valve abnormalities using echocardiography in the prior art.

[0215] It should be noted that the devices and methods disclosed in the embodiments herein can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments herein. In this regard, each block in the flowchart or block diagram may represent a module, program, or part of the code, and the module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0216] The above description is only a description of the preferred embodiments of the present invention and does not limit the scope of the present invention in any way. Any changes and modifications made by those of ordinary skill in the art of the present invention based on the above disclosure belong to the protection scope of the present invention. Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations fall within the scope of the present invention and its equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting the opening distance of the mitral valve, characterized in that Including: Obtain the position information of the mitral valve region of interest according to the acquired current-frame cardiac image; According to the position information of the mitral valve region of interest, use a leaflet segmentation model to segment the mitral valve region of interest corresponding to the current-frame cardiac image to obtain a mitral valve leaflet mask image; Perform connected component analysis on the mitral valve leaflet mask image. If the analysis result of the connected component is that there are two connected components with pixel areas greater than a first preset threshold in the mitral valve leaflet mask image, then according to these two connected components with pixel areas greater than the first preset threshold, extract two leaflet contours, and according to the coordinates of each pixel point on the two leaflet contours, obtain the mitral valve opening distance corresponding to the current-frame cardiac image; The obtaining the mitral valve opening distance corresponding to the current-frame cardiac image according to the coordinates of each pixel point on the two leaflet contours includes: Calculate the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours; According to the minimum pixel distance and the pre-acquired correspondence between the pixel distance and the physical distance, obtain the mitral valve opening distance corresponding to the current-frame cardiac image.

2. The mitral valve opening distance detection method according to claim 1, wherein The calculating the minimum pixel distance between the two leaflet contours according to the coordinates of each pixel point on the two leaflet contours includes: Step A: According to the coordinates of each pixel point on the two leaflet contours, respectively determine a pixel point on each of the two leaflet contours as the starting point; Step B: Make a circle with the two determined starting points as the two endpoints of a diameter; Step C: Determine whether the made circle has new intersections with the two leaflet contours. If not, execute Step D. If so, execute Step E; Step D: Take the diameter of the made circle as the minimum pixel distance between the two leaflet contours; Step E: Determine whether the number of the new intersections is greater than or equal to 2. If so, execute Step E1. If not, execute Step E2; Step E1: Make a circle with two new intersections located on different leaflet contours as the two endpoints of a new diameter, and return to execute Step C; Step E2: Make a circle with the new intersection and the original intersection located on another leaflet contour as the two endpoints of a new diameter, and return to execute Step C.

3. The mitral valve opening distance detection method according to claim 2, characterized in that For Step E1, the method further includes: Determine whether the absolute value of the difference between the distance between two new intersections located on different leaflet contours and the diameter of the currently made circle is less than a second preset threshold; If so, take the distance between the two new intersections as the minimum pixel distance between the two leaflet contours; If not, make a circle with the two new intersections as the two endpoints of a new diameter; For Step E2, the method further includes: Determine whether the absolute value of the difference between the distance between the new intersection and the original intersection located on another leaflet contour and the diameter of the currently made circle is less than the second preset threshold; If so, the distance between the new intersection point and the original intersection point located on the other leaflet contour is taken as the minimum pixel distance between the two leaflet contours; If not, a circle is drawn with the new intersection point and the original intersection point located on the other leaflet contour as the two endpoints of a new diameter.

4. The mitral valve opening distance detection method according to claim 2, characterized in that, The determining of a pixel point as the starting point on each of the two leaflet contours according to the coordinates of the respective pixel points on the two leaflet contours includes: For each leaflet contour, according to the coordinates of the respective pixel points on the leaflet contour, the pixel point located on the leftmost side is taken as the starting point on the leaflet contour; or For each leaflet contour, according to the coordinates of the respective pixel points on the leaflet contour, the pixel point with the minimum sum of the X coordinate and the Y coordinate is taken as the starting point on the leaflet contour.

5. The mitral valve opening distance detection method according to claim 1, wherein The calculating of the minimum pixel distance between the two leaflet contours according to the coordinates of the respective pixel points on the two leaflet contours includes: According to the coordinates of the respective pixel points on the two leaflet contours, a decision boundary for distinguishing the two leaflet contours and support vectors located on each of the leaflet contours are determined, and the support vector is the pixel point on the leaflet contour that is closest to the decision boundary; For each support vector, a perpendicular line is drawn from the support vector to the decision boundary. If the perpendicular line intersects with the other leaflet contour, the support vector is taken as the target support vector; For each target support vector, the pixel distance between the target support vector and the corresponding intersection point is calculated; The calculated minimum pixel distance is taken as the minimum pixel distance between the two leaflet contours.

6. The mitral valve opening distance detection method according to claim 5, wherein The determining of the decision vectors on each of the leaflet contours and the support vectors on each of the leaflet contours includes: The respective pixel points on the two leaflet contours are binary-classified to obtain the pixel points on the upper leaflet contour and the pixel points on the lower leaflet contour, so as to determine the decision boundary; The pixel point on the upper leaflet contour that is closest to the decision boundary is taken as the support vector on the upper leaflet contour, and the pixel point on the lower leaflet contour that is closest to the decision boundary is taken as the support vector on the lower leaflet contour.

7. The mitral valve opening distance detection method according to claim 1, wherein After obtaining the mitral valve opening intervals corresponding to all frames of cardiac motion images, the method further includes: Drawing the two leaflet contours, the mitral valve opening interval diameter line, and the text content of the maximum mitral valve opening interval on the cardiac motion image corresponding to the maximum mitral valve opening interval and outputting.

8. The mitral valve opening distance detection method according to claim 1, wherein The method further includes: Drawing an opening interval spectrogram for characterizing the corresponding relationship between the number of frames and the mitral valve opening interval according to the time sequence and the mitral valve opening interval corresponding to each frame of cardiac motion image.

9. The mitral valve opening distance detection method according to claim 1, wherein, The obtaining of the position information of the mitral valve region of interest according to the acquired current frame of cardiac motion image includes: Using a target detection model to detect the current frame of cardiac motion image to obtain the position information of the mitral valve region of interest.

10. An electronic device, characterized in that, It includes a processor and a memory, and a computer program is stored on the memory. When the computer program is executed by the processor, it implements the mitral valve opening distance detection method described in any one of claims 1 to 9.

11. A readable storage medium, characterized in that, A computer program is stored in the readable storage medium. When the computer program is executed by a processor, it implements the mitral valve opening distance detection method described in any one of claims 1 to 9.

Images