Myocardial infarction region segmentation method and system based on image processing
Patent Information
- Application Number
- CN202411773715.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-05
- Publication Date
- 2026-10-09
AI Technical Summary
[0003]本发明提供基于图像处理的心肌梗死区域分割方法及系统，以解决现有的问题
[0039]本发明实施例中，获取心脏CT灰度图像中的若干个连通域，将每个连通域的相邻连通域记为参考连通域，根据每个连通域内所有像素点的灰度值和每个连通域对应的参考连通域的数量，得到每个连通域的灰度差值，其根据像素点灰度值的分析，确定每个连通域的灰度值差异，提高了心肌梗死区域区分的准确性。对所有连通域进行合并得到若干个目标区域；根据目标区域内每个连通域中任意两个像素点之间的距离，得到每个连通域的目标直线；根据目标区域内连通域的目标直线之间的夹角，从所有连通域中筛选出若干个目标连通域，其根据目标区域内每个连通域中像素点距离的分析，确定出目标区域中方向一致的连通域的数量，保障了心肌梗死区域区分的准确性。根据每个目标区域内所有连通域的灰度差值和目标连通域的数量，得到每个目标区域的显著程度，其根据综合考虑目标区域内多个连通域的特征，避免了根据单一特征对心肌梗死区域进行区分的缺点。根据每个目标区域的显著程度，从所有目标局域中筛选出若干个高关注区域；根据每个高关注区域的显著程度和高关注区域内所有像素点灰度值，得到每个高关注区域内所有像素点的更新灰度值，其根据高关注区域内像素点的灰度值和连通域的灰度差异的分析，对像素点的灰度值进行调整，进一步保障了心肌梗死区域区分的准确性。根据每个高关注区域内所有像素点的更新灰度值，得到心肌梗死区域。至此本发明通过对心脏CT灰度图像进行划分，计算划分的每个区域的显著程度，准确分割出心肌梗死区域，提高了基于图像处理的心肌梗死区域分割的准确性。
Smart Images
Figure CN122887930A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image processing technology, and more specifically to a method and system for segmenting myocardial infarction regions based on image processing. Background Technology
[0002] Cardiac CT scans are an imaging technique that uses computer technology to generate detailed images of the heart and its surrounding structures, which can be used to diagnose myocardial infarction (heart attack). Myocardial infarction is a serious heart condition involving localized necrosis of myocardial tissue. In cardiac CT scans, the infarcted area typically exhibits specific morphological and grayscale characteristics that differ from other normal or diseased tissues. Therefore, to more accurately identify and analyze the infarcted area, it needs to be segmented from the entire cardiac CT scan image. In existing techniques, considering the insignificant difference between the infarcted and normal areas in cardiac CT scans, the CA saliency detection algorithm is used to calculate the saliency of each region in the cardiac CT scan image. Different regions are then enhanced to varying degrees based on their saliency. However, existing CA saliency detection algorithms often fail to effectively distinguish the infarcted area from the myocardial infarction area when processing cardiac CT scans because some normal regions have higher grayscale values. This results in these normal regions receiving higher saliency values using traditional saliency detection algorithms, making it difficult to effectively differentiate them from the infarcted area and thus reducing the segmentation accuracy of the infarcted area in cardiac CT scans. Summary of the Invention
[0003] This invention provides a method and system for segmenting myocardial infarction regions based on image processing, in order to solve existing problems.
[0004] The image processing-based myocardial infarction region segmentation method and system of the present invention adopts the following technical solution:
[0005] One embodiment of the present invention provides a method for segmenting myocardial infarction regions based on image processing, the method comprising the following steps:
[0006] Obtain several connected components from a grayscale image of a heart CT scan;
[0007] The adjacent connected components of each connected component are denoted as reference connected components. The grayscale difference of each connected component is obtained based on the grayscale values of all pixels in each connected component and the number of reference connected components corresponding to each connected component.
[0008] Merge all connected components to obtain several target regions; based on the distance between any two pixels in each connected component within the target region, obtain the target line of each connected component; based on the angle between the target lines of the connected components within the target region, select several target connected components from all connected components.
[0009] The saliency of each target region is obtained based on the gray-scale difference of all connected components within each target region and the number of target connected components.
[0010] Based on the salience of each target region, several high-interest regions are selected from all target local areas; based on the salience of each high-interest region and the grayscale values of all pixels within each high-interest region, the updated grayscale values of all pixels within each high-interest region are obtained.
[0011] The myocardial infarction region is obtained by updating the grayscale values of all pixels within each high-interest region.
[0012] Furthermore, the specific steps for obtaining the grayscale difference value of each connected component based on the grayscale values of all pixels within each connected component and the number of reference connected components corresponding to each connected component are as follows:
[0013] The average gray value of all pixels in each connected component is recorded as the gray threshold of each connected component.
[0014] The absolute value of the difference between the grayscale threshold of each connected component and the corresponding reference connected component is denoted as the grayscale difference between each connected component and the corresponding reference connected component.
[0015] In the grayscale differences between each connected component and all corresponding reference connected components, the number of other reference connected components whose grayscale differences between each connected component and any one of its corresponding reference connected components are greater than the grayscale differences between each connected component and other corresponding reference connected components is denoted as the threshold number of the number of reference connected components between each connected component and the corresponding any one of its corresponding reference connected components.
[0016] The grayscale difference of each connected component is obtained based on the grayscale threshold of each connected component and its corresponding reference connected component, as well as the number threshold of each connected component and its corresponding reference connected component.
[0017] Furthermore, the specific formula for calculating the grayscale difference of each connected component, based on the grayscale threshold of each connected component and its corresponding reference connected component, and the quantity threshold of each connected component and its corresponding reference connected component, is as follows:
[0018]
[0019] Among them, b i N represents the gray-level difference of the i-th connected component. i This represents the number of reference connected components corresponding to the i-th connected component. This represents the grayscale threshold of the i-th connected component. Let n represent the grayscale threshold of the m-th reference connected component corresponding to the i-th connected component. i,mLet || represent the threshold number of connected components of the i-th connected component and the corresponding m-th reference connected component, where || is the absolute value function.
[0020] Furthermore, the specific steps for obtaining the target straight line of each connected component based on the distance between any two pixels in each connected component within the target region are as follows:
[0021] The straight line connecting the two longest-distance pixels within each connected region of the target area is denoted as the target straight line for each connected region.
[0022] Furthermore, the specific steps for selecting several target connected components from all connected components based on the included angle between target lines within the target region are as follows:
[0023] Within the k-th target region, the average of the minimum angles between the i-th connected component and the target lines of all other connected components is denoted as the angle value of the i-th connected component.
[0024] Connected components with an included angle value less than a preset included angle threshold are denoted as target connected components.
[0025] Furthermore, the specific formula for calculating the saliency of each target region based on the grayscale difference of all connected components within each target region and the number of target connected components is as follows:
[0026]
[0027] Among them, s k M represents the saliency of the k-th target region, where K represents the number of target regions in the cardiac CT grayscale image, and M represents the significance of the k-th target region. k M represents the number of connected components in the k-th target region. k,n b represents the number of connected components in the nth target region (excluding the kth target region). k,z b represents the grayscale difference of the z-th connected component within the k-th target region. k,n,j e represents the grayscale difference of the j-th connected component within the n-th target region (excluding the k-th target region). k e represents the number of connected components in the k-th target region. k,n Let || represent the number of connected components in the nth target region (excluding the kth target region), and || represent the absolute value function.
[0028] Furthermore, the specific steps for selecting several high-interest regions from all target regions based on the salience of each target region are as follows:
[0029] The average significance of all target regions is recorded as the standard significance.
[0030] Target regions with a significance level greater than the standard significance level are designated as high-interest regions.
[0031] Furthermore, the specific calculation formula for obtaining the updated grayscale values of all pixels within each high-interest region based on the saliency of each high-interest region and the grayscale values of all pixels within the high-interest region is as follows:
[0032]
[0033] Among them, f x,y s' represents the updated grayscale value of the y-th pixel within the x-th high-interest region. x h represents the significance of the x-th region of high interest. x,y b' represents the grayscale value of the y-th pixel within the x-th region of high interest. x,y This represents the gray-level difference of the connected components containing the y-th pixel within the x-th region of high interest. `norm()` is the floor function, and `norm()` is the linear normalization function.
[0034] Furthermore, the specific steps for obtaining the myocardial infarction region based on the updated grayscale values of all pixels within each high-interest region are as follows:
[0035] In the cardiac CT grayscale image, the grayscale value of each pixel in each region of high interest is replaced with the updated grayscale value to obtain an enhanced cardiac CT image;
[0036] The myocardial infarction region in the enhanced image of the heart CT was obtained by segmenting using a trained segmentation neural network.
[0037] The present invention also proposes an image processing-based myocardial infarction region segmentation system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned image processing-based myocardial infarction region segmentation method.
[0038] The beneficial effects of the technical solution of the present invention are:
[0039] In this embodiment of the invention, several connected components are obtained from a grayscale image of a heart CT scan. Adjacent connected components of each component are designated as reference connected components. The grayscale difference of each connected component is obtained based on the grayscale values of all pixels within that component and the number of reference connected components. This analysis of pixel grayscale values determines the grayscale difference between each connected component, improving the accuracy of myocardial infarction region differentiation. All connected components are merged to obtain several target regions. The target straight line of each connected component is obtained based on the distance between any two pixels within each target region. Several target connected components are selected from all connected components based on the angle between the target straight lines within each target region. The number of connected components with consistent orientation within each target region is determined based on the analysis of pixel distances within that target region, ensuring the accuracy of myocardial infarction region differentiation. The saliency of each target region is obtained based on the grayscale difference of all connected components within each target region and the number of target connected components. This comprehensive consideration of the characteristics of multiple connected components within the target region avoids the drawback of differentiating myocardial infarction regions based on a single feature. Based on the saliency of each target region, several high-interest regions are selected from all target local areas. Based on the saliency of each high-interest region and the grayscale values of all pixels within that region, updated grayscale values are obtained for all pixels within each high-interest region. Adjustments to the pixel grayscale values are made based on the analysis of grayscale differences between pixels within the high-interest region and connected components, further ensuring the accuracy of myocardial infarction region segmentation. The myocardial infarction region is then obtained based on the updated grayscale values of all pixels within each high-interest region. Thus, this invention, by dividing a cardiac CT grayscale image and calculating the saliency of each segmented region, accurately segments the myocardial infarction region, improving the accuracy of image processing-based myocardial infarction region segmentation. Attached Figure Description
[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0041] Figure 1 This is a flowchart illustrating the steps of the image processing-based myocardial infarction region segmentation method of the present invention. Detailed Implementation
[0042] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the image processing-based myocardial infarction region segmentation method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0044] The following description, in conjunction with the accompanying drawings, details the specific scheme of the image processing-based myocardial infarction region segmentation method and system provided by this invention.
[0045] Please see Figure 1 The diagram illustrates a flowchart of an image processing-based myocardial infarction region segmentation method according to an embodiment of the present invention, which includes the following steps:
[0046] Step S001: Obtain several connected components in the grayscale image of the heart CT.
[0047] Acquire cardiac CT scan images, convert the cardiac CT scan images to grayscale, and obtain cardiac CT grayscale images.
[0048] Several connected regions in a grayscale cardiac CT image were obtained using a region growing algorithm.
[0049] It should be noted that grayscale conversion and region growing algorithms are well-known techniques, and the specific methods will not be introduced here.
[0050] Step S002: Denote the adjacent connected components of each connected component as reference connected components. Based on the gray values of all pixels in each connected component and the number of reference connected components corresponding to each connected component, obtain the gray value difference of each connected component.
[0051] Myocardial infarction not only leads to myocardial tissue fibrosis and necrosis, but is also often accompanied by inflammation. The infarcted area is lighter in color and is often surrounded by normal areas. This results in larger gray values and more concentrated connected regions for pixels in the myocardial infarction area in cardiac CT grayscale images. In contrast, while there are also areas with larger gray values in the normal area, the connected regions of the normal area are usually isolated, and the connected regions of the normal area are surrounded by areas with smaller gray values. The gray value difference within each connected region is obtained by calculating the difference between the mean gray value of all pixels in each connected region and the mean gray value of all pixels in the adjacent connected regions.
[0052] The adjacent connected components of each connected component are denoted as reference connected components;
[0053] The average gray value of all pixels in each connected component is recorded as the gray threshold of each connected component.
[0054] The absolute value of the difference between the grayscale threshold of each connected component and the corresponding reference connected component is denoted as the grayscale difference between each connected component and the corresponding reference connected component.
[0055] In the grayscale differences between each connected component and all corresponding reference connected components, the number of other reference connected components whose grayscale differences between each connected component and any one of its corresponding reference connected components are greater than the grayscale differences between each connected component and other corresponding reference connected components is denoted as the threshold number of the number of reference connected components between each connected component and the corresponding any one of its corresponding reference connected components.
[0056] Taking the i-th connected component as an example, the formula for calculating the gray-level difference of the i-th connected component is:
[0057]
[0058] Among them, b i N represents the gray-level difference of the i-th connected component. i This represents the number of reference connected components corresponding to the i-th connected component. This represents the grayscale threshold of the i-th connected component. Let n represent the grayscale threshold of the m-th reference connected component corresponding to the i-th connected component. i,m Let || represent the threshold number of connected components of the i-th connected component and the corresponding m-th reference connected component, where || is the absolute value function.
[0059] It should be noted that: This represents the grayscale difference between the i-th connected component and the corresponding m-th reference connected component. The larger the grayscale difference, the greater the difference in the average grayscale value of the i-th connected component and the surrounding connected components.
[0060] Following the above method, this embodiment can obtain the grayscale difference of each connected domain.
[0061] Step S003: Merge all connected components to obtain several target regions; based on the distance between any two pixels in each connected component within the target region, obtain the target line of each connected component; based on the angle between the target lines of the connected components within the target region, select several target connected components from all connected components.
[0062] The connected domains in the myocardial infarction area are clustered, and the cellular tissue in the myocardial infarction area is relatively dense, so the connected domains in the myocardial infarction area are relatively close together.
[0063] The centroid distance is preset to l. Based on the preset centroid distance, the CA saliency detection algorithm is used to merge all connected components to obtain several target regions.
[0064] It should be noted that: l is the preset centroid distance in this embodiment, which is 10, and this is used as an example for description. The CA saliency detection algorithm is a well-known technique, and its specific method will not be described here.
[0065] Because the connected domains within a myocardial infarction region are relatively concentrated, they are easier to classify into a single region. Therefore, the greater the difference between target regions, the greater the significance of the target region, and the greater the likelihood that the target region is a myocardial infarction region. Furthermore, in myocardial infarction regions, cell fibrosis is generally accompanied by inflammation, which disrupts the texture characteristics of cell growth in myocardial tissue. The proliferation and differentiation of normal cells in the heart have a certain directionality, that is, the connected domains within a normal region have a certain directionality. However, due to cell aggregation, cell fibrosis, and inflammation, the cell proliferation and differentiation in myocardial infarction regions exhibit disordered and chaotic characteristics, that is, the connected domains within a myocardial infarction region have no obvious directionality.
[0066] The straight line connecting the two longest-distance pixels within each connected component of the target region is denoted as the target straight line for each connected component.
[0067] Within the k-th target region, the average of the minimum angles between the i-th connected component and the target lines of all other connected components is denoted as the angle value of the i-th connected component.
[0068] Following the above method, this embodiment can obtain the included angle value of each connected component within the k-th target region.
[0069] A preset angle threshold is used to define connected components whose angle values are less than the preset angle threshold as target connected components.
[0070] It should be noted that the preset included angle threshold in this embodiment is 5°, and this will be used as an example for description. The more connected components there are within the target area, the more consistent the directions of the connected components within the target area are, and the higher the probability that the target area belongs to the myocardial infarction area.
[0071] Step S004: Based on the grayscale difference of all connected components within each target region and the number of target connected components, obtain the saliency of each target region.
[0072] Taking the k-th target region as an example, the formula for calculating the salience of the k-th target region is:
[0073]
[0074] Among them, s k M represents the saliency of the k-th target region, where K represents the number of target regions in the cardiac CT grayscale image, and M represents the significance of the k-th target region. k M represents the number of connected components in the k-th target region. k,n b represents the number of connected components in the nth target region (excluding the kth target region). k,z b represents the grayscale difference of the z-th connected component within the k-th target region. k,n,j e represents the grayscale difference of the j-th connected component within the n-th target region (excluding the k-th target region). k e represents the number of connected components in the k-th target region. k,n Let || represent the number of connected components in the nth target region (excluding the kth target region), and || represent the absolute value function.
[0075] It should be noted that: This represents the sum of the grayscale differences of all connected components in the k-th target region. The larger the sum, the greater the feature intensity of the k-th target region. Similarly... This represents the sum of grayscale differences of all connected components in the nth target region (excluding the kth target region). The larger the sum, the greater the feature degree of the nth target region (excluding the kth target region).
[0076] Because the number of connected components in different target regions is not the same, the difference in the number of connected components in different target regions may lead to the same cumulative gray-level difference in connected components. Therefore, by using the number of target connected components in the target region, different weighting coefficients are assigned to different target regions. The more target connected components in a target region, the greater the probability that the target region is a normal myocardial tissue region. Conversely, the less the target region is, the less likely it is to be a normal myocardial tissue region, that is, the greater the probability that the target region is a myocardial infarction region. This represents the significant difference between the k-th target region and other target regions at a centroid distance of l, i.e., the significance of the k-th target region.
[0077] Following the above method, this embodiment can obtain the salience of each target region.
[0078] Step S005: Based on the salience of each target region, select several high-interest regions from all target local areas; based on the salience of each high-interest region and the grayscale values of all pixels within the high-interest region, obtain the updated grayscale values of all pixels within each high-interest region.
[0079] At a scale where the centroid distance is preset to l, the obtained target area may contain normal areas, making the diagnosis less accurate.
[0080] The average of the saliency of all target regions is denoted as the standard saliency.
[0081] Target regions with a significance level greater than the standard significance level are designated as high-interest regions.
[0082] Taking the x-th high-interest region as an example, the formula for calculating the updated grayscale value of the y-th pixel within the x-th high-interest region is:
[0083]
[0084] Among them, f x,y s' represents the updated grayscale value of the y-th pixel within the x-th high-interest region. x h represents the significance of the x-th region of high interest. x,y b' represents the grayscale value of the y-th pixel within the x-th region of high interest. x,y This represents the gray-level difference of the connected components containing the y-th pixel within the x-th region of high interest. `norm()` is the floor function, and `norm()` is the linear normalization function.
[0085] It should be noted that: s' x ×b' x,y This indicates the degree of enhancement of the grayscale value of the y-th pixel within the x-th high-interest region. The greater the enhancement, the greater the likelihood that the y-th pixel belongs to the myocardial infarction region; the smaller the enhancement, the greater the likelihood that the y-th pixel belongs to the normal region.
[0086] Following the above method, this embodiment can obtain the updated grayscale values of all pixels in each high-interest area.
[0087] Step S006: Obtain the myocardial infarction region based on the updated grayscale values of all pixels within each high-interest region.
[0088] In a cardiac CT grayscale image, the grayscale value of each pixel in each region of high interest is replaced with the updated grayscale value to obtain an enhanced cardiac CT image.
[0089] The trained segmentation neural network was used to segment the myocardial infarction region in the enhanced images of cardiac CT.
[0090] It should be noted that the embodiments of the present invention use a segmentation neural network to identify normal areas and myocardial infarction areas in the enhanced images of segmented cardiac CT.
[0091] The relevant content of segmentation neural networks is as follows:
[0092] The segmentation neural network used in this embodiment is the Mask R-CNN neural network; the dataset used is an enhanced image dataset of cardiac CT scans. Mask R-CNN is a well-known technology, and its specific method will not be described here. The full Chinese name of Mask R-CNN is “Mask Region-based Convolutional Neural Network”, and its full English name is “Mask Region-based Convolutional Neural Network”.
[0093] The pixels that need to be segmented are divided into two categories. The labeling process for the corresponding labels in the training set is as follows: for single-channel semantic labels, pixels belonging to the normal region are labeled as 0, and pixels belonging to the myocardial infarction region are labeled as 1.
[0094] The network's task is classification, so the loss function used is the cross-entropy loss function.
[0095] The process of obtaining normal and myocardial infarction areas from enhanced cardiac CT images using a segmentation neural network is a well-known technique, and the specific method will not be described here.
[0096] This invention is now complete.
[0097] In summary, in this embodiment of the invention, several connected components in a cardiac CT grayscale image are obtained. Each connected component's adjacent connected components are designated as reference connected components. The grayscale difference of each connected component is obtained based on the average grayscale value of all pixels within that component and the number of reference connected components corresponding to that component. A preset centroid distance is used. All connected components are merged based on this preset centroid distance to obtain several target regions. The target straight line of each connected component is obtained based on the distance between any two pixels within each connected component of the target region. Several target connected components are selected from all connected components based on the angle between the target straight lines of the connected components within the target region. The saliency of each target region is obtained based on the grayscale difference of all connected components within each target region and the number of target connected components. Based on the saliency of each target region, several high-interest regions are selected from all target regions. The updated grayscale values of all pixels within each high-interest region are obtained based on the saliency of each high-interest region and the grayscale values of all pixels within that high-interest region. Finally, the myocardial infarction region is obtained based on the updated grayscale values of all pixels within each high-interest region. This invention improves the accuracy of myocardial infarction region segmentation based on image processing by dividing grayscale images of the heart into segments and calculating the salience of each segment.
[0098] The present invention also provides an image processing-based myocardial infarction region segmentation system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned image processing-based myocardial infarction region segmentation method.
[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for segmenting myocardial infarction regions based on image processing, characterized in that, The method includes the following steps: Obtain several connected components from a grayscale image of a heart CT scan; The adjacent connected components of each connected component are denoted as reference connected components. The grayscale difference of each connected component is obtained based on the grayscale values of all pixels in each connected component and the number of reference connected components corresponding to each connected component. Merge all connected components to obtain several target regions; based on the distance between any two pixels in each connected component within the target region, obtain the target line of each connected component; based on the angle between the target lines of the connected components within the target region, select several target connected components from all connected components. The saliency of each target region is obtained based on the gray-scale difference of all connected components within each target region and the number of target connected components. Based on the salience of each target region, several high-interest regions are selected from all target local areas; based on the salience of each high-interest region and the grayscale values of all pixels within each high-interest region, the updated grayscale values of all pixels within each high-interest region are obtained. The myocardial infarction region is obtained by updating the grayscale values of all pixels within each high-interest region.
2. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific steps for obtaining the grayscale difference value of each connected component based on the grayscale values of all pixels within each connected component and the number of reference connected components corresponding to each connected component are as follows: The average gray value of all pixels in each connected component is recorded as the gray threshold of each connected component. The absolute value of the difference between the grayscale threshold of each connected component and the corresponding reference connected component is denoted as the grayscale difference between each connected component and the corresponding reference connected component. In the grayscale differences between each connected component and all corresponding reference connected components, the number of other reference connected components whose grayscale differences between each connected component and any one of its corresponding reference connected components are greater than the grayscale differences between each connected component and other corresponding reference connected components is denoted as the threshold number of the number of reference connected components between each connected component and the corresponding any one of its corresponding reference connected components. The grayscale difference of each connected component is obtained based on the grayscale threshold of each connected component and its corresponding reference connected component, as well as the number threshold of each connected component and its corresponding reference connected component.
3. The image processing-based myocardial infarction region segmentation method according to claim 2, characterized in that, The specific formula for calculating the grayscale difference of each connected component, based on the grayscale threshold of each connected component and its corresponding reference connected component, and the quantity threshold of each connected component and its corresponding reference connected component, is as follows: Among them, b i N represents the gray-level difference of the i-th connected component. i This represents the number of reference connected components corresponding to the i-th connected component. This represents the grayscale threshold of the i-th connected component. Let n represent the grayscale threshold of the m-th reference connected component corresponding to the i-th connected component. i，m Let || represent the threshold number of connected components of the i-th connected component and the corresponding m-th reference connected component, where || is the absolute value function.
4. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific steps for obtaining the target straight line for each connected component based on the distance between any two pixels in each connected component within the target region are as follows: The straight line connecting the two longest-distance pixels within each connected region of the target area is denoted as the target straight line for each connected region.
5. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific steps for selecting several target connected components from all connected components based on the included angle between target lines within the target region are as follows: Within the k-th target region, the average of the minimum angles between the i-th connected component and the target lines of all other connected components is denoted as the angle value of the i-th connected component. Connected components with an included angle value less than a preset included angle threshold are denoted as target connected components.
6. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific formula for calculating the saliency of each target region based on the gray-level difference of all connected components within each target region and the number of target connected components is as follows: Among them, s k M represents the saliency of the k-th target region, where K represents the number of target regions in the cardiac CT grayscale image, and M represents the significance of the k-th target region. k M represents the number of connected components in the k-th target region. k,n b represents the number of connected components in the nth target region (excluding the kth target region). k,z b represents the grayscale difference of the z-th connected component within the k-th target region. k,n,j e represents the grayscale difference of the j-th connected component within the n-th target region (excluding the k-th target region). k e represents the number of connected components in the k-th target region. k,n Let || represent the number of connected components in the nth target region (excluding the kth target region), and || is the absolute value function.
7. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific steps for selecting several high-interest regions from all target local areas based on the salience of each target region are as follows: The average significance of all target regions is recorded as the standard significance. Target regions with a significance level greater than the standard significance level are designated as high-interest regions.
8. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific calculation formula for obtaining the updated grayscale value of all pixels in each high-attention region based on the saliency of each high-attention region and the grayscale values of all pixels in the high-attention region is as follows: Among them, f x,y s' represents the updated grayscale value of the y-th pixel within the x-th high-interest region. x h represents the significance of the x-th region of high interest. x,y b' represents the grayscale value of the y-th pixel within the x-th region of high interest. x,y This represents the grayscale difference of the connected components containing the y-th pixel within the x-th region of high interest. `norm()` is the floor function, and `norm()` is the linear normalization function.
9. The image processing-based myocardial infarction region segmentation method according to claim 1, characterized in that, The specific steps for obtaining the myocardial infarction region based on the updated grayscale values of all pixels within each high-interest region are as follows: In the cardiac CT grayscale image, the grayscale value of each pixel in each region of high interest is replaced with the updated grayscale value to obtain an enhanced cardiac CT image; The myocardial infarction region in the enhanced image of the heart CT was obtained by segmenting using a trained segmentation neural network.
10. A myocardial infarction region segmentation system based on image processing, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the image processing-based myocardial infarction region segmentation method as described in any one of claims 1-9.