Image detection methods, devices, and computer equipment for the mitral valve

Abstract

This application relates to an image detection method, apparatus, computer device, computer-readable storage medium, and computer program product for the mitral valve. The method includes: acquiring multiple frames of images including the heart; performing multi-scale feature extraction on the multiple frames to obtain multi-scale image features for each frame; for each frame, determining the mitral valve region of the heart in each frame based on the multi-scale image features; determining the area of ​​the mitral valve orifice in each frame based on the mitral valve region; and generating a detection result of the mitral valve based on the area of ​​the mitral valve orifice in each frame. This method not only solves the problems of time-consuming and subjective factors that traditional manual measurement methods suffer from, but also significantly improves the efficiency and accuracy of mitral valve assessment, reduces the professional requirements for operators, and enhances the flexibility and practicality of clinical applications.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to an image detection method and device for mitral valve, computer equipment, computer readable storage medium and computer program product. BACKGROUND

[0002] The mitral valve is of great importance for heart analysis. As one of the four valves of the heart, the mitral valve is located between the left atrium and the left ventricle, composed of an anterior leaflet and a posterior leaflet, attached to the annulus or annulus, and connected to the papillary muscle on the left ventricular wall by chordae tendineae. The main function is to open during the filling phase of the heart to allow blood to flow from the left atrium to the left ventricle, and to close during the systole of the heart to prevent blood from flowing back to the left atrium. Accurate assessment of the state and function of the mitral valve is crucial for the diagnosis and treatment of heart disease.

[0003] However, the traditional measurement method for the mitral orifice mainly relies on the experience and manual operation of echocardiogram experts, which is not only time-consuming but also subject to subjective factors, resulting in differences between different observers. The dynamic and real-time ultrasound scanning puts very high requirements on the operator, and the high difficulty of echocardiogram scanning also makes the acquisition and diagnosis process highly subjective and greatly influenced by the operator. Therefore, the use of echocardiogram requires users to pay a very high learning cost. In addition, the manual measurement process is complex, requiring accurate selection of the maximum mitral orifice area during diastole, and repeated operation in multiple cardiac cycles to improve measurement consistency, which is not always widely adopted in clinical practice. SUMMARY

[0004] Therefore, it is necessary to provide an image detection method, device, computer equipment, computer readable storage medium and computer program product for mitral valve, which can improve the measurement efficiency and accuracy of the maximum mitral orifice area.

[0005] In a first aspect, the present application provides an image detection method for mitral valve, comprising:

[0006] obtaining a plurality of images including a heart, performing multi-scale feature extraction on the plurality of images to obtain multi-scale image features of each image;

[0007] For each image, determining the mitral valve region of the heart in each image according to the multi-scale image features of the image;

[0008] determining the area of the mitral orifice in each image based on the mitral valve region of the heart in each image;

[0009] generate a detection result of the mitral valve of the heart according to the area of the mitral orifice in each frame of image.

[0010] In one of the embodiments, the determining, for each frame of image, the mitral valve region of the heart in each frame of image according to the multi-scale image features of the image includes:

[0011] generating a detection box in the image according to the multi-scale image features of the image, the detection box being used to represent the position of different tissue structures in the heart; determining the image quality of the image according to the image quality evaluation rule and the detection box generated in the image; and determining the mitral valve region of the heart in each frame of image according to the image quality of each frame of image.

[0012] In one of the embodiments, the detection box at least includes a left ventricle detection box; and the determining, according to the image quality of each frame of image, the mitral valve region of the heart in each frame of image includes:

[0013] screening, according to the image quality of each frame of image, a plurality of target images from the plurality of frames of image, the plurality of target images being a plurality of continuous frames of image; cropping, for each target image, the target image according to the left ventricle detection box in the target image to obtain a left ventricle image; classifying each pixel point in the left ventricle image to obtain a position type of each pixel point, the position type including belonging to the mitral valve and not belonging to the mitral valve; and segmenting the left ventricle image according to the pixel points with the position type of belonging to the mitral valve to obtain a mitral valve image, the mitral valve image representing the mitral valve region of the heart.

[0014] In one of the embodiments, each frame of image is each frame of mitral valve image; and the generating, according to the area of the mitral orifice in each frame of image, a detection result of the mitral valve of the heart includes:

[0015] converting the area of the mitral orifice in each frame of mitral valve image into a coordinate position in a preset coordinate system; generating a mitral orifice area change curve according to the coordinate position of each frame of mitral valve image in the preset coordinate system and the time stamp of each frame of mitral valve image, the mitral orifice area change curve representing the change of the area of the mitral orifice with time in real time; and determining the detection result of the mitral valve of the heart according to the mitral orifice area change curve.

[0016] In one of the embodiments, the generating, according to the coordinate position of each frame of mitral valve image in the preset coordinate system and the time stamp of each frame of mitral valve image, a mitral orifice area change curve includes:

[0017] According to the coordinate position of each frame of mitral valve image in the preset coordinate system and the time stamp of each frame of mitral valve image, an initial time domain curve is generated; the initial time domain curve is converted from a time domain form to a frequency domain form to obtain an initial frequency domain curve; the initial frequency domain curve is subjected to low-frequency filtering processing through a low-pass filter to obtain a filtered frequency domain curve; and the filtered frequency domain curve is converted from a frequency domain form to a time domain form to obtain a mitral orifice area change curve.

[0018] In one of the embodiments, the determining the detection result of the mitral valve of the heart according to the mitral orifice area change curve comprises:

[0019] According to the mitral orifice area change curve, a maximum area of the mitral orifice is determined; the maximum area of the mitral orifice is compared with a mitral valve stenosis threshold value, and a stenosis detection result of the mitral valve of the heart is determined according to the comparison result; the mitral orifice area change rate curve is generated according to the mitral orifice area change curve, the mitral orifice area change rate curve real-time representing a change trend of the area of the mitral orifice with time; and a recovery detection result of the mitral valve of the heart is determined according to the mitral orifice area change rate curve.

[0020] In a second aspect, the present application further provides an image detection device for a mitral valve, comprising:

[0021] The extraction module is configured to acquire a plurality of images comprising a heart, perform multi-scale feature extraction on the plurality of images, and obtain multi-scale image features of each image;

[0022] The first determination module is configured to, for each image, determine a mitral valve region of the heart in the image according to the multi-scale image features of the image;

[0023] The second determination module is configured to determine an area of a mitral orifice in the image based on the mitral valve region of the heart in the image;

[0024] The generation module is configured to generate a detection result of the mitral valve of the heart according to the area of the mitral orifice in the image.

[0025] In a third aspect, the present application further provides a computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0026] The computer device is configured to acquire a plurality of images comprising a heart, perform multi-scale feature extraction on the plurality of images, and obtain multi-scale image features of each image;

[0027] For each frame of image, according to the multi-scale image features of the frame of image, determine the mitral valve region of the heart in the frame of image;

[0028] Based on the mitral valve region of the heart in the frame of image, determine the area of the mitral orifice in the frame of image;

[0029] According to the area of the mitral orifice in the frame of image, generate the detection result of the mitral valve of the heart.

[0030] In a fourth aspect, the present application further provides a computer readable storage medium, which has a computer program stored thereon, and the computer program is executed by a processor to implement the following steps:

[0031] Obtain a plurality of frames of images including a heart, perform multi-scale feature extraction on the plurality of frames of images, and obtain multi-scale image features of each frame of image;

[0032] For each frame of image, according to the multi-scale image features of the frame of image, determine the mitral valve region of the heart in the frame of image;

[0033] Based on the mitral valve region of the heart in the frame of image, determine the area of the mitral orifice in the frame of image;

[0034] According to the area of the mitral orifice in the frame of image, generate the detection result of the mitral valve of the heart.

[0035] In a fifth aspect, the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the following steps:

[0036] Obtain a plurality of frames of images including a heart, perform multi-scale feature extraction on the plurality of frames of images, and obtain multi-scale image features of each frame of image;

[0037] For each frame of image, according to the multi-scale image features of the frame of image, determine the mitral valve region of the heart in the frame of image;

[0038] Based on the mitral valve region of the heart in the frame of image, determine the area of the mitral orifice in the frame of image;

[0039] According to the area of the mitral orifice in the frame of image, generate the detection result of the mitral valve of the heart.

[0040] The image detection method, device, computer equipment, computer readable storage medium and computer program product for the mitral valve described above, by acquiring multiple frames of images including a heart, performing multi-scale feature extraction on the multiple frames of images, obtaining multi-scale image features of each frame of image, and capturing image features at different scales, the heart structure can be more accurately identified. This approach helps to improve the accuracy of subsequent steps in positioning the heart anatomy (such as the mitral valve). And reduces the high dependence on the experience and skills of the operator, so that even inexperienced operators can obtain high-quality heart images.

[0041] For each frame of image, according to the multi-scale image features of the image, determine the mitral valve region of the heart in each frame of image; automatically detect and locate the mitral valve region without human intervention, greatly improve the work efficiency, and avoid human error. And determine the mitral valve region based on multi-scale features can better adapt to changes and noise in the image, and improve the processing capability of the system for images of different quality and angle.

[0042] Based on the mitral valve region of the heart in each frame of image, determine the area of the mitral orifice in each frame of image; automatically calculate the area of the mitral orifice (MVA) in each frame of image, realize the quantitative evaluation of the function of the mitral valve, and provide more objective data support.

[0043] According to the area of the mitral orifice in each frame of image, generate the detection result of the mitral valve of the heart. By summarizing the MVA data in each frame of image, an MVA curve changing with time can be constructed, which can be used to monitor the functional state of the mitral valve in real time and provide feedback information to the clinician in time.

[0044] Through the above steps, not only the problems such as time-consuming and subjective influence of traditional manual measurement method are solved, but also the efficiency and accuracy of mitral valve evaluation are significantly improved, the professional requirements for the operator are reduced, and the flexibility and practicality of clinical application are enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other related drawings without creative labor.

[0046] Figure 1 An application environment diagram for the image detection method for the mitral valve in an embodiment;

[0047] Figure 2A flowchart of an image detection method for a mitral valve in an embodiment;

[0048] Figure 3 A framework diagram of an image detection method for a mitral valve in an embodiment;

[0049] Figure 4 A flowchart of an image detection method for a mitral valve in another embodiment;

[0050] Figure 5 A block diagram of an image detection device for a mitral valve in an embodiment;

[0051] Figure 6 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0052] In order to make the purposes, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0053] The image detection method for a mitral valve provided by the embodiments of the present application can be applied in an application environment as shown in Figure 1 . The terminal 102 communicates with the server 104 through a network. The data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The terminal 102 is used to generate an image detection request for a mitral valve, and send the image detection request for the mitral valve to the server 104, so that the server 104 generates a detection result of a mitral valve of a heart according to an area of a mitral orifice in each frame of image. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle device, a projection device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The head-mounted device can be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc. The server 104 can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0054] In an exemplary embodiment, as shown in Figure 2 , an image detection method for a mitral valve is provided. The method is applied in Figure 1The following steps 202 to 208 are described by taking the server 104 in the system 100 as an example. Among them:

[0055] At step 202, a plurality of images including a heart are obtained, and multi-scale feature extraction is performed on the plurality of images to obtain multi-scale image features of each frame of image.

[0056] The plurality of images refer to a series of static pictures extracted from a continuous video stream or dynamic image. Each frame represents a picture at a certain moment in the video. For echocardiograms, these frames capture the state of the heart at different time points, including diastole (when the heart relaxes) and systole (when the heart pumps blood).

[0057] The multi-scale image features refer to representative information in the image extracted through processing under different resolutions or different size windows. For example, only the general outline of the entire heart can be seen at a larger scale; while more detailed structural details, such as the specific morphology of the mitral valve, can be seen at a smaller scale.

[0058] Specifically, first, a plurality of images including a heart are obtained, including: converting a dynamic image of the heart from an ultrasound device (especially a parasternal short-axis mitral valve view PSAX-MV perspective) into a single frame of image. The central region of each frame of image is cropped to ensure that the image focuses on the heart structure and removes unnecessary background information. The cropped image is adjusted to a uniform resolution (e.g., 320x320 pixels) by using a bilinear interpolation method to obtain a plurality of images of the heart, so as to ensure consistency and efficiency of subsequent processing.

[0059] Then, multi-scale feature extraction is performed on the plurality of images, including: creating a set of input images containing multiple resolution versions. Each version corresponds to a different scale, which is used to capture different levels of features from global to local. A mitral valve image detection network is used, which adopts a feature pyramid architecture and can extract multi-scale image features at different scales. Two subnets of the mitral valve image detection network perform classification tasks (identify key anatomical structures) and regression tasks (optimize the position and size of the bounding box) respectively, to accurately mark the key anatomical structures of the heart, including the left ventricle, the right ventricle, the anterior leaflet of the mitral valve, and the posterior leaflet of the mitral valve.

[0060] Finally, multi-scale image features of each frame of image are obtained, including: after processing by the feature pyramid and the mitral valve image detection network, a plurality of multi-scale feature maps corresponding to each frame of image are obtained. These feature maps contain information at different levels, including both the overall outline of the heart and the detailed parts such as the mitral valve.

[0061] Step 204, for each frame of image, determine the mitral valve region of the heart in each frame of image according to the multi-scale image features of the image.

[0062] The mitral valve is one of the four valves of the heart, located between the left atrium and the left ventricle. It is composed of two leaflets: the anterior leaflet and the posterior leaflet. The two leaflets are attached to a ring-shaped structure called the annulus or annulus, and are connected to the papillary muscles on the left ventricular wall by chordae tendineae.

[0063] The mitral valve region refers to the part in the heart image that can clearly identify the anterior and posterior leaflets of the mitral valve and the surrounding anatomical structures. Specifically, the mitral valve region refers to the specific anatomical location containing the mitral valve that can be identified and defined in each frame of echocardiogram image.

[0064] Specifically, after completing the anatomical structure detection, the mitral valve region in each frame of image is further accurately defined. Specifically, it includes: inputting the multi-frame image after secondary cropping and resolution adjustment into the segmentation branch model. The segmentation branch model is specifically used to accurately segment the mitral leaflets according to the multi-scale image features of the multi-frame image, so as to clearly define the mitral valve region. The segmentation model will output a binary image in which the specific position of the mitral valve is marked. This region is the mitral valve region, which contains the anterior and posterior leaflets of the mitral valve and the range of their attachment points.

[0065] Step 206, based on the mitral valve region of the heart in each frame of image, determine the area of the mitral orifice in each frame of image.

[0066] The mitral orifice refers to the opening through which blood flows from the left atrium to the left ventricle when the mitral valve is open. This opening is formed by the two leaflets of the mitral valve, namely the anterior leaflet and the posterior leaflet, and is open during the filling period (diastolic period) of the heart and closed during the contraction period (systolic period).

[0067] The area of the mitral orifice refers to the maximum cross-sectional area of the mitral orifice in the open state. It is an important parameter for measuring the size of the mitral orifice, and has important significance in evaluating diseases such as mitral stenosis.

[0068] Specifically, first, the mitral valve orifice contour is outlined, including: based on the segmentation result, using edge detection algorithm or morphological operation to identify and outline the contour of the mitral valve orifice. Optionally, special attention is paid to the image at the end of diastole, because at this time the mitral valve orifice reaches the maximum open state. The outlined contour is smoothed to remove possible noise points or abnormal protrusions, and finally to ensure that the obtained mitral valve orifice contour is continuous and smooth. Then calculate the area of the mitral valve orifice, including: using appropriate geometric method or numerical integration technique, according to the contour point coordinates to calculate the area of the mitral valve orifice, and record the calculated area of the mitral valve orifice. The optional calculation method includes but is not limited to polygon area formula, green formula, etc.

[0069] Step 208, according to the area of the mitral valve orifice in each frame of image, the detection result of the mitral valve of the heart is generated.

[0070] Among them, the detection result refers to the information set that can reflect the functional state of the heart mitral valve generated by the image detection method. It integrates various features extracted from multiple images and is obtained through a series of processing steps (such as feature extraction, region definition, area calculation, etc.), aiming to provide reliable diagnostic basis for doctors.

[0071] Specifically, first, the mitral valve orifice area data is summarized, including: extracting the calculated mitral valve orifice area value from each video frame. And arrange these mitral valve orifice area values in time sequence to form a time series data set to reflect the change of mitral valve orifice area with heart cycle.

[0072] Then build and optimize the mitral valve orifice area change curve, including: based on the time series data, draw the frame-by-frame mitral valve orifice area change curve. This curve reflects the change of mitral valve orifice area with time in real time, especially in diastole and systole. Optionally, in order to improve the accuracy of detection, a denoising module based on Fourier transform is introduced. The specific steps include fast Fourier transform, low-pass filtering and inverse fast Fourier transform, to remove high-frequency noise and retain low-frequency components, so as to obtain more refined and smooth mitral valve orifice area change curve.

[0073] Next, quantitatively evaluate the function of the mitral valve, including: by observing the mitral valve orifice area change curve, identify whether the mitral valve opening is restricted. When the mitral valve orifice area decreases to a certain threshold (for example, less than 2 square centimeters for moderate stenosis, and less than 1.5 square centimeters for severe stenosis), it is suggested that there may be mitral stenosis. The mitral valve orifice area change curve can help evaluate the closure of the mitral valve in systole. If the systolic mitral valve orifice area is found to be abnormally increased, it may be a sign of mitral regurgitation.

[0074] Finally, qualitative description and diagnosis suggestions are provided, including: combining the mitral valve orifice area change curve and the quantitative evaluation results, the functional state of the mitral valve is described in detail, such as whether there is stenosis or regurgitation phenomenon, and the severity thereof. For patients who have received treatment, the change of the mitral valve function before and after treatment is tracked by using the mitral valve orifice area change curve, and the effectiveness of the treatment is evaluated. According to the evaluation results, diagnosis suggestions are provided for doctors, including the selection of further examination and treatment plan, etc.

[0075] In one embodiment, based on the multi-scale image features of the image, a detection box is generated in the image, which is used to represent the location of different tissue structures in the heart; according to the image quality evaluation rules and the detection box generated in the image, the image quality of the image is determined; according to the image quality of each frame of image, the mitral valve region of the heart in each frame of image is determined.

[0076] Wherein, the detection box refers to a rectangular boundary in the image used to mark or surround a specific target object. In this application, the detection box is used to represent the location of different tissue structures in the heart, such as left ventricle, right ventricle, anterior leaflet of mitral valve and posterior leaflet of mitral valve. Tissue structure refers to different components inside the heart, which are manifested as different morphological and textural features in echocardiogram images, including but not limited to left ventricle, right ventricle, anterior leaflet of mitral valve and posterior leaflet of mitral valve, etc. Image quality evaluation rules refer to a set of predefined standards or criteria used to evaluate the quality of each frame of image. These rules are based on whether the key anatomical structures can be clearly displayed in the image, as well as the overall clarity, contrast and other factors of the image. Image quality refers to the extent to which the image meets the predetermined quality standards, especially in capturing and displaying the internal structures of the heart.

[0077] Specifically, referring to Figure 3The detection branch in the mitral valve image detection network first identifies and locates key anatomical structures (such as the left ventricle, the right ventricle, the anterior leaflet of the mitral valve, and the posterior leaflet of the mitral valve) according to the multi-scale image features of each frame of image by the classification subnet and the regression subnet in the mitral valve image detection network, and outputs detection boxes containing the accurate boundary boxes of these structures. Then a set of image quality evaluation rules is established. Optionally, when the detection box cannot identify the left ventricle, the image quality is 1 point, when the detection box can identify the left ventricle but cannot identify the right ventricle, the image quality is 2 points, when the detection box can identify the left ventricle and the right ventricle but cannot identify the anterior leaflet and the posterior leaflet of the valve, the image quality is 3 points, when the detection box can identify the left ventricle and the right ventricle but can only identify one valve, the image quality is 4 points, and when the detection box can identify the left ventricle, the right ventricle, the anterior leaflet and the posterior leaflet of the valve at the same time, the image quality is 5 points. Combined with the detection box information, an image quality score is generated for each frame of image according to the image quality evaluation rules. Then record the video segment with a quality score greater than a preset quality threshold (for example, 3), to ensure that the images used for further analysis are of high quality. Finally, based on the image quality score, the images used for further analysis are selected from each frame of image, and the image quality of the selected images used for further analysis is determined.

[0078] Due to the introduction of the image quality evaluation rules, only high-quality images are used for further analysis. This method reduces the risk of false positives or false negatives caused by low-quality images (such as the presence of noise or artifacts), enhancing the robustness of the system and the reliability of the results. Moreover, by scoring the quality of each frame of image, high-quality image segments suitable for further analysis can be quickly screened out, thereby speeding up the overall diagnostic process and improving clinical efficiency.

[0079] In one embodiment, a plurality of target images are selected from the plurality of images according to the image quality of each frame of image, the plurality of target images being a plurality of consecutive frames of images; for each target image, the target image is cropped according to the left ventricular detection box in the target image to obtain a left ventricular image; each pixel point in the left ventricular image is classified to obtain a position type of each pixel point, the position type including belonging to the mitral valve and not belonging to the mitral valve; the left ventricular image is segmented according to the pixel points belonging to the mitral valve to obtain a mitral valve image, the mitral valve image representing a mitral valve region of the heart.

[0080] In one embodiment, a plurality of target images are selected from the plurality of images according to the image quality of each frame of image, the plurality of target images being a plurality of consecutive frames of images; for each target image, the target image is cropped according to the left ventricular detection box in the target image to obtain a left ventricular image; each pixel point in the left ventricular image is classified to obtain a position type of each pixel point, the position type including belonging to the mitral valve and not belonging to the mitral valve; the left ventricular image is segmented according to the pixel points belonging to the mitral valve to obtain a mitral valve image, the mitral valve image representing a mitral valve region of the heart.

[0081] The left ventricle bounding box refers to a rectangular boundary box generated in each target image, which is used to mark and surround the position of the left ventricle. The bounding box is identified and located in the left ventricle of the heart by the mitral valve image detection network.

[0082] The left ventricle image refers to the image obtained by cropping from each target image according to the left ventricle bounding box. The cropping process removes unnecessary background information, making the image more focused on the left ventricle and the surrounding area.

[0083] The position type refers to the classification result of each pixel point in the left ventricle image, which is divided into two types, including: belonging to the mitral valve (indicating that the pixel point is located in the mitral valve area) and not belonging to the mitral valve (indicating that the pixel point is not in the mitral valve area).

[0084] The mitral valve image refers to the image obtained by segmentation processing, which specifically represents the mitral valve area of the heart. The image is composed of all pixel points classified as "belonging to the mitral valve", clearly identifying the positions of the anterior and posterior leaflets of the mitral valve.

[0085] Specifically, referring to the segmentation branch in Figure 3 , first, high-quality and continuous target images are selected from the original multi-frame images, including: based on pre-defined image quality evaluation rules, a quality score is generated for each image. And record the video segment with a quality score greater than the pre-set quality threshold, to ensure that these images are of high quality and continuous. This step ensures that the images used for subsequent analysis are clear and reliable. Understandably, by selecting multiple high-quality target images based on image quality scores, it is ensured that the images used for subsequent analysis are clear and reliable. This reduces the risk of false positives or false negatives caused by low-quality images, improving the accuracy of the final analysis results.

[0086] Then generate the left ventricle bounding box in each target image, and crop the image according to the box to obtain an image focused on the left ventricle and its surrounding area, including: using the mitral valve image detection network to process each target image, the network uses a feature pyramid architecture to extract multi-scale image features and outputs an accurate bounding box containing the position of the left ventricle, i.e. the detection box. According to the identified left ventricle bounding box, the ultrasound image is cropped again to remove unnecessary background information, making the image more focused on the left ventricle and the surrounding key anatomical structures.

[0087] Next, classify each pixel point in the cropped left ventricle image to determine whether it belongs to the mitral valve area, including: using a pixel classification model to classify each pixel point in the left ventricle image into two position types: "belongs to the mitral valve" and "does not belong to the mitral valve". A binary image is generated, which marks the position type of each pixel point, especially the specific position of the mitral valve.

[0088] Finally, according to the pixel classification results, the left ventricular image is segmented to obtain an image specifically representing the mitral valve region, including: based on the classification results, all pixel points classified as "belonging to the mitral valve" are combined together to form a mitral valve image. This image clearly identifies the positions of the anterior leaflet and posterior leaflet of the mitral valve. Optionally, additional smoothing processing or morphological operations are introduced to remove possible noise points or abnormal protrusions, ensuring the continuity and smoothness of the segmentation results.

[0089] Since the target image is cropped twice using the left ventricular detection frame, focusing on the key anatomical structures inside the heart, especially the mitral valve. This method enhances the ability to capture details such as the mitral valve, ensuring the accuracy of subsequent analysis. The entire process from image screening to the final definition of the mitral valve region is automated and does not require human intervention. This not only improves work efficiency, but also maintains consistency and repeatability, unaffected by operator subjective factors, enhancing the robustness and reliability of the system. And by classifying each pixel point in the left ventricular image and segmenting according to the pixel points belonging to the mitral valve, an image specifically representing the mitral valve region is generated. This detailed segmentation and classification helps to more comprehensively evaluate the functional status of the mitral valve, providing more objective and reliable diagnostic basis.

[0090] In one embodiment, the area of the mitral valve orifice in each frame of the mitral valve image is converted into a coordinate position in a preset coordinate system; according to the coordinate position of each frame of the mitral valve image in the preset coordinate system and the time stamp of each frame of the mitral valve image, a mitral valve orifice area change curve is generated, which real-time represents the change of the area of the mitral valve orifice with time; according to the mitral valve orifice area change curve, the detection result of the mitral valve of the heart is determined.

[0091] Among them, the preset coordinate system refers to a unified reference frame for standardizing the mitral valve orifice area information in different frames of images. This coordinate system can be two-dimensional or three-dimensional, and has a fixed origin and axial direction, ensuring that all data are compared and analyzed under the same standard.

[0092] The coordinate position refers to the specific position of the area of the mitral valve orifice in each frame of the mitral valve image in the preset coordinate system. These positions can be described by one or more pairs of numerical values (such as x, y coordinates), depending on the type of coordinate system used. Understandably, each coordinate position represents the spatial positioning of the area of the mitral valve orifice in a frame of image, so that the data between different frames can be accurately compared and tracked, especially when evaluating the change of the mitral valve orifice over time.

[0093] The timestamp refers to a time marker that records the moment of capturing each frame of image. It can be an absolute time (e.g., specific date and time) or a relative time (e.g., time seconds relative to the start of the video). In this application, the timestamp provides information on the time dimension for each frame of image, so that the changes in the mitral valve orifice area can be tracked in the time sequence. This is crucial for constructing an accurate mitral valve orifice area change curve, as it allows the area changes to be associated with specific cardiac cycle phases.

[0094] The mitral valve orifice area change curve refers to a curve that can represent the changes in the mitral valve orifice area over time in real time, generated by aggregating the mitral valve orifice area data in multiple frames of image and combining their corresponding coordinate positions and timestamps.

[0095] Specifically, first, the calculated mitral valve orifice area in each frame of mitral valve image is converted to a unified preset coordinate system to ensure that all data are compared under the same standard, including: establishing a preset coordinate system (e.g., two-dimensional or three-dimensional coordinate system) with a fixed origin and axis. For each frame of mitral valve image, the position information of the mitral valve orifice area is converted to the specific coordinate position in the preset coordinate system. This step may involve geometric transformation, scaling, etc. to ensure the consistency of data between different frames.

[0096] Then, the timestamp of each frame of mitral valve image is recorded, including: adding a timestamp to each frame of image to record the specific time of capture or the time point relative to the start of the video. This can be achieved through the time coding function of the ultrasound device itself, or manually added in the post-processing.

[0097] Next, based on the coordinate position of each frame of mitral valve image in the preset coordinate system and its timestamp, the mitral valve orifice area data in all frames is aggregated to generate a curve that can represent the changes in the mitral valve orifice area over time in real time, including: collecting the mitral valve orifice area values in all frames of image and arranging them into a time series data set together with the corresponding timestamps. Based on the time series data, a frame-by-frame mitral valve orifice area change curve is drawn. This curve intuitively reflects the changing trend of the mitral valve orifice area in different cardiac cycles, especially in the diastolic and systolic periods. Optionally, to improve the accuracy of prediction, a denoising module based on Fourier transform (such as fast Fourier transform, low-pass filtering, and inverse fast Fourier transform) is introduced to optimize the original curve to remove high-frequency noise and retain low-frequency components, thereby obtaining a more refined and smooth mitral valve orifice area change curve.

[0098] Finally, based on the generated mitral valve orifice area change curve, the functional status of the heart's mitral valve is comprehensively evaluated, and the final detection results are obtained, including: by observing the mitral valve orifice area change curve, identifying whether the mitral valve opening is restricted (for example, to evaluate mitral valve stenosis), or evaluating the closure of the mitral valve in the systolic phase (for example, to find evidence of mitral valve regurgitation). In combination with the mitral valve orifice area change curve and the quantitative evaluation results, the functional status of the mitral valve is described in detail, such as whether there is stenosis or regurgitation and the severity thereof. For patients who have received treatment, the change in mitral valve function before and after treatment is tracked using the mitral valve orifice area change curve to evaluate the effectiveness of the treatment.

[0099] By converting the mitral valve orifice area into a preset coordinate system, all measurement data are ensured to be in the same standard framework. This eliminates errors caused by different scanning angles, device resolution differences, and other factors, improving the consistency and reliability of data analysis. The introduction of time stamps gives each frame of image clear time information, allowing accurate tracking of the change in mitral valve orifice area over time. This is very important for understanding the dynamic behavior of the heart over a complete cardiac cycle, especially for evaluating the different performances of the mitral valve in the diastolic and systolic phases. The generated mitral valve orifice area change curve can visually show the trend of the change in mitral valve orifice area over time, helping to identify potential problems such as mitral valve stenosis or regurgitation. After denoising and optimization of the curve by Fourier transform technology, the obtained mitral valve orifice area change curve is more refined and smooth, further improving the stability and accuracy of the prediction, thereby improving the reliability of the diagnosis.

[0100] In one embodiment, an initial time-domain curve is generated based on the coordinate position of each frame of mitral valve image in the preset coordinate system and the time stamp of each frame of mitral valve image; the initial time-domain curve is converted from time-domain form to frequency-domain form to obtain an initial frequency-domain curve; the initial frequency-domain curve is subjected to low-frequency filtering processing by a low-pass filter to obtain a filtered frequency-domain curve; and the filtered frequency-domain curve is converted from frequency-domain form to time-domain form to obtain a mitral valve orifice area change curve.

[0101] The initial time-domain curve refers to a curve that can represent the change in mitral valve orifice area over time in real time, which is generated by aggregating the coordinate positions of each frame of mitral valve image in the preset coordinate system and their corresponding time stamps.

[0102] Time domain form refers to the way signals or data are arranged in a time sequence. In this form, each data point is associated with a specific time point, and it is often used to directly show how the signal changes over time. Understandably, time domain data is easy to understand and visualize, and the specific trend of mitral valve orifice area change over time can be directly observed. However, time domain data may contain noise and other high-frequency fluctuations, affecting the accuracy of analysis.

[0103] Frequency domain form refers to the way signals or data are represented by mathematical transformation (such as Fourier transform) from time domain to frequency components. In the frequency domain, the signal is decomposed into a combination of sinusoidal waves of different frequencies and amplitudes. Understandably, the frequency domain form allows us to analyze the frequency composition of the signal, identifying which frequency components represent actual physiological phenomena and which are noise. This is very helpful for removing high-frequency noise and retaining low-frequency useful information.

[0104] Initial frequency domain curve refers to the curve obtained by transforming the initial time domain curve from time domain to frequency domain through fast Fourier transform. This curve shows the frequency component distribution of the original mitral valve orifice area change curve.

[0105] Low-pass filter refers to a signal processing tool that retains low-frequency components in the signal while removing high-frequency components. It determines which frequency components can be retained by setting a cutoff frequency. In this application, the low-pass filter removes high-frequency noise from the initial frequency domain curve and only retains low-frequency components, which are more consistent with actual physiological phenomena, thereby improving the accuracy of subsequent analysis.

[0106] Filtered frequency domain curve refers to the frequency domain curve after being processed by the low-pass filter. This curve has removed most of the high-frequency noise and only retains useful low-frequency components.

[0107] Specifically, referring to the correction branch in Figure 3 , first, based on the coordinate position of each frame of mitral valve image in the preset coordinate system and its timestamp, the mitral valve orifice area data in all frames are summarized to construct an initial time domain curve. This curve directly reflects the trend of mitral valve orifice area change over time. Then, using fast Fourier transform, the initial time domain curve is transformed from time domain to frequency domain to obtain an initial frequency domain curve. This step reveals the frequency component distribution of the original signal, including low-frequency and high-frequency components. Next, a low-pass filter is applied to the initial frequency domain curve to remove high-frequency noise and retain only low-frequency components by selecting an appropriate cutoff frequency. This ensures that the final curve better reflects the actual mitral valve orifice area change. Finally, through inverse fast Fourier transform, the filtered frequency domain curve is converted back to time domain form to obtain an optimized mitral valve orifice area change curve. This curve is more refined and smooth, enhancing the quality of the signal and improving the stability and accuracy of the prediction.

[0108] Since the generated initial time-domain curve intuitively reflects the trend of the mitral valve orifice area changing over time, especially in different phases of the cardiac cycle (such as diastole and systole). This method can accurately capture the dynamic changes of the mitral valve orifice area, providing detailed physiological information, which helps to more accurately evaluate the heart function. And by Fourier transform, the initial time-domain curve is converted into a frequency domain form, and then a low-pass filter is applied to remove high-frequency noise and only keep low-frequency components. This step significantly improves the quality of the signal, removing high-frequency fluctuations and artifacts in the original data, making the final mitral valve orifice area change curve smoother and more stable. And the whole process from generating the initial time-domain curve to finally generating the optimized mitral valve orifice area change curve is automated, reducing the need for human intervention and reducing the influence of operator's subjective factors. This not only improves work efficiency, but also ensures the consistency and repeatability of the results.

[0109] In one embodiment, according to the mitral valve orifice area change curve, the maximum area of the mitral valve orifice is determined; the maximum area of the mitral valve orifice is compared with the mitral valve stenosis threshold, and according to the comparison result, the stenosis detection result of the mitral valve of the heart is determined; according to the mitral valve orifice area change curve, the mitral valve orifice area change rate curve is generated, which real-time represents the trend of the area of the mitral valve orifice changing over time; according to the mitral valve orifice area change rate curve, the recovery detection result of the mitral valve of the heart is determined.

[0110] Among them, the maximum area of the mitral valve orifice refers to the maximum value of the mitral valve orifice area in the mitral valve orifice area change curve. It usually appears at the end of diastole, that is, when the left ventricle is filled. Understandably, the maximum area of the mitral valve orifice is a key parameter for evaluating the functional status of the mitral valve, which is used to identify whether there is mitral stenosis and other problems. Under normal circumstances, the mitral valve orifice area is about 4-6 square centimeters; when MVA decreases to 2 square centimeters or less, it is considered to be moderate stenosis; less than 1.5 square centimeters is considered to be severe stenosis.

[0111] The mitral valve stenosis threshold refers to the standard critical value for judging whether the mitral valve is stenotic. It can be set to a specific area value, such as 2 square centimeters or 1.5 square centimeters, to distinguish between normal and abnormal conditions.

[0112] The comparison result refers to the result obtained by comparing the maximum area of the mitral valve orifice with the mitral valve stenosis threshold. Through this comparison, it can be determined whether the mitral valve orifice area is below the critical value, thereby judging whether there is a stenosis phenomenon.

[0113] The stenosis detection result refers to the specific conclusion of mitral valve stenosis determined based on the comparison result. It may include information such as whether there is stenosis, the degree of stenosis (mild, moderate or severe), etc.

[0114] The mitral valve orifice area change rate curve refers to a curve generated by taking the derivative of the mitral valve orifice area change curve, which can represent the trend of the mitral valve orifice area changing over time in real time. This curve reflects the speed of area change. Understandably, the mitral valve orifice area change rate curve not only shows the trend of the mitral valve orifice area change, but also reveals the time points of rapid change, which is very important for evaluating the functional status of the mitral valve. In addition, it helps to identify potential problems such as mitral regurgitation and evaluate the treatment effect.

[0115] The recovery detection result refers to the specific conclusion of heart mitral valve recovery determined based on the mitral valve orifice area change rate curve. This includes but is not limited to evaluating the changes in mitral valve function before and after treatment, and monitoring the progress of the patient's recovery process.

[0116] Specifically, first find the maximum area of the mitral valve orifice from the mitral valve orifice area change curve, including: by analyzing the mitral valve orifice area change curve, identify the maximum value. This maximum value usually occurs at the end of diastole, i.e. when the left ventricle is filling. And clearly record the specific value of the maximum area as the basis for subsequent comparison.

[0117] Then compare the maximum area of the mitral valve orifice with the preset stenosis threshold (e.g. 2 square centimeters or 1.5 square centimeters) to determine whether there is stenosis, including: refer to medical standards, set the critical value of mitral valve stenosis (e.g. less than 2 square centimeters for moderate stenosis, less than 1.5 square centimeters for severe stenosis). Compare the maximum area determined in the previous step with the preset stenosis threshold to determine whether it is below the critical value. And clearly record the comparison result, such as the maximum area being greater than 2 square centimeters or the maximum area being less than 1.5 square centimeters.

[0118] Then according to the comparison result, the specific stenosis detection conclusion is obtained, including whether there is stenosis and its severity, including: based on the comparison result, generate a detailed stenosis detection report. For example, if the maximum area is less than 1.5 square centimeters, it is determined as severe stenosis; if it is between 1.5 and 2 square centimeters, it is determined as moderate stenosis.

[0119] Next, by differentiating the mitral valve orifice area change curve, a curve that can represent the real-time trend of mitral valve orifice area change over time is generated. This includes: differentiating the mitral valve orifice area change curve, calculating the rate of change of area at each moment, and generating the mitral valve orifice area change rate curve. Optionally, to improve the quality of the mitral valve orifice area change rate curve, smoothing or filtering techniques can be introduced to remove possible noise points or abnormal protrusions, thereby ensuring the continuity and smoothness of the mitral valve orifice area change rate curve.

[0120] Finally, based on the mitral valve orifice area change rate curve, the changes in mitral valve function before and after treatment were assessed, the patient's recovery process was monitored, and corresponding medical advice was provided, including: assessing the treatment effect by observing the trend of the mitral valve orifice area change rate curve before and after treatment. For example, if the curve tends to stabilize after treatment and the maximum area increases, it indicates that the treatment is effective.

[0121] By determining the maximum area of ​​the mitral valve orifice and comparing it with a stenosis threshold, the presence and severity of mitral stenosis can be accurately determined. This method provides clear quantitative evidence and reduces errors caused by subjective judgment. The generated mitral valve orifice area change rate curve can characterize the trend of mitral valve orifice area change over time in real time, helping doctors dynamically monitor changes in cardiac function, especially the cardiac recovery process before and after treatment. By taking the derivative of the mitral valve orifice area change curve and generating the change rate curve, noise and abnormal fluctuations in the original data can be better removed, preserving true functional change information and improving signal quality and reliability.

[0122] In one exemplary embodiment, such as Figure 4 As shown, steps 402 to 408 are included. Wherein:

[0123] Step 402: Acquire multiple frames of images including the heart, and perform multi-scale feature extraction on the multiple frames to obtain multi-scale image features for each frame.

[0124] In step 404, a detection box is generated in the image based on the image features of the multi-scale image, and the detection box is used to represent the position of different tissue structures in the heart; the image quality of the image is determined according to the image quality evaluation rule and the detection box generated in the image; the target image is screened out from the multiple images according to the image quality of each image, and the target image is a plurality of continuous images; for each target image, the target image is cropped according to the left ventricular detection box in the target image to obtain a left ventricular image; each pixel point in the left ventricular image is classified to obtain the position type of each pixel point, and the position type includes belonging to the mitral valve and not belonging to the mitral valve; the left ventricular image is segmented according to the pixel point belonging to the mitral valve to obtain a mitral valve image, and the mitral valve image represents the mitral valve region of the heart.

[0125] In step 406, the area of the mitral orifice in each image is determined based on the mitral valve region of the heart in each image.

[0126] In step 408, the area of the mitral orifice in each mitral valve image is converted into a coordinate position in a preset coordinate system; an initial time-domain curve is generated according to the coordinate position of each mitral valve image in the preset coordinate system and the time stamp of each mitral valve image; the initial time-domain curve is converted from the time-domain form to the frequency-domain form to obtain an initial frequency-domain curve; the initial frequency-domain curve is filtered by a low-pass filter to obtain a filtered frequency-domain curve; the filtered frequency-domain curve is converted from the frequency-domain form to the time-domain form to obtain a mitral orifice area change curve, which represents the change of the area of the mitral orifice with time; the maximum area of the mitral orifice is determined according to the mitral orifice area change curve; the maximum area of the mitral orifice is compared with the mitral valve stenosis threshold, and the stenosis detection result of the mitral valve of the heart is determined according to the comparison result; the mitral orifice area change rate curve is generated according to the mitral orifice area change curve, and the mitral orifice area change rate curve represents the change trend of the area of the mitral orifice with time; the recovery detection result of the mitral valve of the heart is determined according to the mitral orifice area change rate curve.

[0127] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above embodiments can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of the steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0128] Based on the same inventive concept, the embodiments of the present application also provide an image detection device for the mitral valve for implementing the above-mentioned image detection method for the mitral valve. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more image detection device embodiments for the mitral valve provided below can refer to the limitations of the image detection method for the mitral valve in the above, which will not be repeated here.

[0129] In one exemplary embodiment, as shown in Figure 5 An image detection device 500 for the mitral valve is provided, comprising: an extraction module 502, a first determination module 504, a second determination module 506, and a generation module 508, wherein:

[0130] The extraction module 502 is configured to obtain a plurality of images including a heart, perform multi-scale feature extraction on the plurality of images, and obtain multi-scale image features of each image;

[0131] The first determination module 504 is configured to, for each image, determine a mitral valve region of the heart for each image according to the multi-scale image features of the image;

[0132] The second determination module 506 is configured to determine the area of the mitral valve orifice in each image based on the mitral valve region of the heart for each image;

[0133] The generation module 508 is configured to generate a detection result of the mitral valve of the heart according to the area of the mitral valve orifice in each image.

[0134] In one of the embodiments, the first determining module 504 is configured to generate a detection frame in the image based on the image features of multiple scales of the image, the detection frame being used to represent the positions of different tissue structures in the heart; determine the image quality of the image according to the image quality evaluation rule and the detection frame generated in the image; and determine the mitral valve region of the heart in each frame of image according to the image quality of each frame of image.

[0135] In one of the embodiments, the first determining module 504 is configured to filter out a plurality of target images from the plurality of images according to the image quality of each frame of image, the plurality of target images being a plurality of continuous frames of image; for each target image, perform cropping on the target image according to the left ventricle detection frame in the target image to obtain a left ventricle image; classify each pixel point in the left ventricle image to obtain a position type of each pixel point, the position type including belonging to the mitral valve and not belonging to the mitral valve; and perform segmentation on the left ventricle image according to the pixel points with the position type belonging to the mitral valve to obtain a mitral valve image, the mitral valve image representing the mitral valve region of the heart.

[0136] In one of the embodiments, the first determining module 508 is configured to convert the area of the mitral valve orifice in each frame of mitral valve image into a coordinate position in a preset coordinate system; generate a mitral valve orifice area change curve according to the coordinate position of each frame of mitral valve image in the preset coordinate system and the timestamp of each frame of mitral valve image, the mitral valve orifice area change curve representing the change of the area of the mitral valve orifice over time in real time; and determine the detection result of the mitral valve of the heart according to the mitral valve orifice area change curve.

[0137] In one of the embodiments, the first determining module 508 is configured to generate an initial time-domain curve according to the coordinate position of each frame of mitral valve image in the preset coordinate system and the timestamp of each frame of mitral valve image; convert the initial time-domain curve from the time-domain form to the frequency-domain form to obtain an initial frequency-domain curve; perform low-frequency filtering processing on the initial frequency-domain curve through a low-pass filter to obtain a filtered frequency-domain curve; and convert the filtered frequency-domain curve from the frequency-domain form to the time-domain form to obtain the mitral valve orifice area change curve.

[0138] In one of the embodiments, the first determining module 508 is configured to determine the maximum area of the mitral valve orifice according to the mitral valve orifice area change curve; compare the maximum area of the mitral valve orifice with a mitral valve stenosis threshold value, and determine the stenosis detection result of the mitral valve of the heart according to the comparison result; generate a mitral valve orifice area change rate curve according to the mitral valve orifice area change curve, the mitral valve orifice area change rate curve representing the change trend of the area of the mitral valve orifice over time in real time; and determine the recovery detection result of the mitral valve of the heart according to the mitral valve orifice area change rate curve.

[0139] The various modules in the image detection device for the mitral valve described above can be implemented by software, hardware and combinations thereof, in whole or in part. The various modules described above can be embedded in the processor in the computer device in hardware form or independent of the processor in the computer device, or stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the various modules.

[0140] In an exemplary embodiment, a computer device, which can be a server, is provided, and an internal structure diagram of the computer device can be as shown in Figure 6 The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is configured to store data related to the image detection for the mitral valve. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with terminals outside through network connection. The computer program is executed by the processor to implement an image detection method for the mitral valve.

[0141] Those skilled in the art can understand that Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0142] In an embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in the above method embodiments.

[0143] In an embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program is executed by the processor to implement the steps in the above method embodiments.

[0144] In an embodiment, a computer program product is provided, including a computer program, and the computer program is executed by the processor to implement the steps in the above method embodiments.

[0145] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0146] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. In the embodiments provided in the present application, any reference to memory, database or other medium can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., without being limited thereto.

[0147] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, any combination of these technical features is deemed to be within the scope of the present application.

[0148] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method of image detection for a mitral valve, characterized in that, The method comprises: acquiring a plurality of images including a heart, performing multi-scale feature extraction on the plurality of images to obtain multi-scale image features of each image; for each image, determining a mitral valve image in the image according to the multi-scale image features of the image; the mitral valve image represents a mitral valve region of the heart; performing edge detection on the mitral valve region of the heart in each image to obtain contour points of a mitral orifice, and determining an area of the mitral orifice in each image according to coordinates of the contour points of the mitral orifice by using a polygon area formula; converting the area of the mitral orifice in each image into a coordinate position in a preset coordinate system; adding a time stamp to each image, and generating an initial time-domain curve according to the coordinate position of each image in the preset coordinate system and the time stamp of each image; converting the initial time-domain curve from a time-domain form into a frequency-domain form to obtain an initial frequency-domain curve; performing low-frequency filtering on the initial frequency-domain curve by using a low-pass filter to obtain a filtered frequency-domain curve; converting the filtered frequency-domain curve from a frequency-domain form into a time-domain form to obtain a mitral orifice area change curve; the mitral orifice area change curve represents a change of the area of the mitral orifice with time in real time; determining a maximum area of the mitral orifice according to the mitral orifice area change curve, comparing the maximum area of the mitral orifice with a mitral valve stenosis threshold, and determining a stenosis detection result of the mitral valve of the heart according to a comparison result; performing differential processing on the mitral orifice area change curve to determine a change rate of the area at each time, generating a mitral orifice area change rate curve based on the change rate of the area at each time, and the mitral orifice area change rate curve represents a change trend of the area of the mitral orifice with time in real time; determining a recovery detection result of the mitral valve of the heart according to the mitral orifice area change rate curve.

2. The method of claim 1, wherein, The method comprises: generating a detection frame in the image based on the multi-scale image features of the image; the detection frame is used to represent positions of different tissue structures in the heart; determining an image quality of the image according to an image quality evaluation rule and the detection frame generated in the image; determining the mitral valve image in the image according to the image quality of the image.

3. The method of claim 2, wherein, The detection frame at least comprises a left ventricle detection frame. The method comprises: screening a plurality of target images from the plurality of images according to the image quality of each image; the plurality of target images are images of a plurality of continuous frames; for each target image, performing cropping on the target image according to a left ventricle detection frame in the target image to obtain a left ventricle image; classifying each pixel point in the left ventricle image to obtain a position type of each pixel point; the position type comprises belonging to a mitral valve or not belonging to the mitral valve. According to the position type, pixels belonging to the mitral valve are segmented from the left ventricle image to obtain a mitral valve image.

4. The method of claim 1, wherein, The area of the mitral orifice is a maximum cross-sectional area of the mitral orifice in an open state.

5. An image detection apparatus for a mitral valve, characterized by, The device comprises: An extraction module is configured to acquire a plurality of images including a heart, perform multi-scale feature extraction on the plurality of images, and obtain multi-scale image features of each image; A first determination module is configured to determine, for each image, a mitral valve image in the image according to the multi-scale image features of the image; the mitral valve image represents a mitral valve region of the heart; A second determination module is configured to perform edge detection on the mitral valve region of the heart in each image to obtain contour points of a mitral orifice, and determine an area of the mitral orifice in each mitral valve image according to coordinates of the contour points of the mitral orifice using a polygon area formula; A generation module is configured to convert the area of the mitral orifice in each mitral valve image into a coordinate position in a preset coordinate system, add a time stamp to each mitral valve image, generate an initial time-domain curve according to the coordinate position of each mitral valve image in the preset coordinate system and the time stamp of each mitral valve image, convert the initial time-domain curve from a time-domain form to a frequency-domain form to obtain an initial frequency-domain curve, perform low-frequency filtering on the initial frequency-domain curve through a low-pass filter to obtain a filtered frequency-domain curve, convert the filtered frequency-domain curve from a frequency-domain form to a time-domain form to obtain a mitral orifice area change curve, the mitral orifice area change curve represents a change of the area of the mitral orifice with time in real time, determine a maximum area of the mitral orifice according to the mitral orifice area change curve, compare the maximum area of the mitral orifice with a mitral valve stenosis threshold, and determine a stenosis detection result of the mitral valve of the heart according to a comparison result, perform differential processing on the mitral orifice area change curve to determine a change rate of the area at each time, generate a mitral orifice area change rate curve based on the change rate of the area at each time, the mitral orifice area change rate curve represents a change trend of the area of the mitral orifice with time in real time, and determine a recovery detection result of the mitral valve of the heart according to the mitral orifice area change rate curve.

6. The apparatus of claim 5, wherein, The first determination module generates a detection box in the image according to the multi-scale image features of the image, the detection box is used to represent positions of different tissue structures in the heart, determines an image quality of the image according to an image quality evaluation rule and the detection box generated in the image, and determines the mitral valve image in each image according to the image quality of each image.

7. The apparatus of claim 5, wherein, The area of the mitral orifice is a maximum cross-sectional area of the mitral orifice in an open state. 8.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-7. The processor executes the computer program to implement the steps of the method in any one of claims 1 to 4.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 4.

10. A computer program product comprising a computer program, characterized in that, The computer program, which is executed by a processor, implements the steps of the method according to any one of claims 1 to 4.

Images