A method for coronary ultrasound image atherosclerotic plaque segmentation

Abstract

The present application relates to the technical field of image segmentation, and particularly relates to a method for segmenting atherosclerotic plaque in coronary ultrasound images. After determining a blood vessel segment, the method first determines the abnormality degree of the gray scale distribution of each blood vessel segment according to the feature that adjoining normal blood vessel segments usually exhibit similar gray scale distribution. Further, the method determines the abnormality degree of the diameter of each blood vessel segment according to the feature that the diameter between adjoining normal blood vessel segments usually does not change greatly. Then, the method determines the abnormality degree of the change trend of each blood vessel segment according to the feature that the presence of atherosclerotic plaque region will destroy the similarity of the gray scale distribution of normal blood vessels. Finally, the method screens abnormal blood vessel segments affected by atherosclerotic plaque by comprehensively considering the abnormality degree of the gray scale distribution, the abnormality degree of the diameter and the abnormality degree of the change trend. Thus, the method only performs image segmentation in abnormal blood vessel segments, so that the segmented atherosclerotic plaque region is more accurate.

Description

Technical Field

[0001] This invention relates to the field of image segmentation technology, and specifically to a method for segmenting atherosclerotic plaques in coronary ultrasound images. Background Technology

[0002] Current techniques for segmenting atherosclerotic plaques in coronary ultrasound images typically rely on the high grayscale values ​​of these plaques, directly using the Otsu's method (maximum between-class variance) to segment the vascular region. However, physiological fluctuations in human blood vessels can produce artifacts and virtual images. This can lead to the Otsu's method misinterpreting regions corresponding to these artifacts and virtual images as atherosclerotic plaques. In other words, the accuracy of current techniques for segmenting atherosclerotic plaques directly from the vascular region in coronary ultrasound images using Otsu's method is relatively low. Summary of the Invention

[0003] To address the low accuracy of existing techniques that directly segment atherosclerotic plaque regions from coronary ultrasound images using the Otsu's method, this application aims to provide a method for segmenting atherosclerotic plaques from coronary ultrasound images. The specific technical solution adopted is as follows: The first aspect of this application provides a method for segmenting atherosclerotic plaques in coronary ultrasound images, including: Acquire coronary ultrasound images; perform edge segmentation on the coronary ultrasound images to identify all vessel segments; Based on the region display of each vessel segment during the threshold segmentation process of the coronary ultrasound image, the segmentation threshold of each vessel segment is determined; based on the segmentation threshold deviation between each vessel segment and its adjacent segments, the degree of grayscale distribution abnormality of each vessel segment is determined. The degree of abnormality in the diameter of each blood vessel segment is determined based on the deviation in the width of each blood vessel segment from its adjacent segments; the degree of abnormality in the trend of change of each blood vessel segment is determined based on the deviation in the number of pixels of each blood vessel segment during the threshold segmentation process. Based on the degree of abnormality in grayscale distribution, the degree of abnormality in vessel diameter, and the degree of abnormality in the trend of change, the probability of abnormality for each vessel segment is determined; abnormal vessel segments are screened out based on the probability of abnormality; image segmentation is performed on the abnormal vessel segments to determine the atherosclerotic plaque region.

[0004] Furthermore, the process of obtaining the vascular segment includes: The coronary ultrasound image is subjected to Canny edge detection to determine the corresponding edge lines; the area enclosed by the edge lines with the largest area is taken as the vascular region; the vascular region is subjected to morphological processing to extract the skeleton and determine the corresponding vascular skeleton. On the vascular skeleton, the number of adjacent skeleton pixels of each skeleton pixel is used as the corresponding reference number; the vascular skeleton is divided into at least two skeleton segments using skeleton pixels with a reference number greater than 2 as segmentation points; and the corresponding vascular segment is determined according to the region corresponding to each skeleton segment in the vascular region.

[0005] Furthermore, the process of obtaining the segmentation threshold includes: The minimum gray value of the coronary ultrasound image is used as the starting gray value threshold; the maximum gray value of the coronary ultrasound image is used as the ending gray value threshold; all gray values ​​between the starting gray value threshold and the ending gray value threshold are traversed with a preset gray value interval as the traversal step size to determine all gray value thresholds. In the process of traversing and segmenting coronary ultrasound images in ascending order of grayscale threshold, the grayscale threshold of the first segmented region in each vessel segment is used as the first boundary threshold of each vessel segment; the grayscale threshold of the first complete segmentation of each vessel segment is used as the second boundary threshold of each vessel segment; and the segmentation threshold of each vessel segment is determined based on the average value between the first boundary threshold and the second boundary threshold.

[0006] Furthermore, the process of obtaining the degree of grayscale distribution anomaly includes: Each vascular segment is sequentially designated as the target segment; all vascular segments adjacent to the target segment are designated as the corresponding reference segments; the threshold deviation value of each reference segment is determined based on the difference between the segmentation threshold of the target segment and the segmentation threshold of each reference segment; the mean of the segmentation threshold differences between each reference segment and all other reference segments is negatively correlated and normalized to determine the reference weight of each reference segment. The weighted deviation value of each reference segment is determined by multiplying the reference weight by the threshold deviation value; the mean of the weighted deviation values ​​of all reference segments corresponding to the target segment is normalized to determine the degree of grayscale distribution anomaly of the target segment.

[0007] Furthermore, the process of obtaining the degree of abnormality in the blood vessel diameter includes: The corresponding blood vessel diameter is determined based on the overall relative distance between the two edge lines on both sides of the vascular skeleton of each blood vessel segment; the reference diameter of each blood vessel segment is determined based on the average blood vessel diameter of all adjacent blood vessel segments; the difference between the blood vessel diameter of each blood vessel segment and the reference diameter is normalized to determine the degree of abnormality of the blood vessel diameter of each blood vessel segment.

[0008] Furthermore, the process of obtaining the blood vessel diameter includes: In each blood vessel segment, the two edge lines on both sides of the blood vessel skeleton are taken as the blood vessel edge lines; the minimum distance between each pixel on each blood vessel edge line and all pixels on the other blood vessel edge line is taken as the local distance of each pixel on each blood vessel edge line; the blood vessel diameter of each blood vessel segment is determined based on the average of the local distances of all pixels on the two blood vessel edge lines of each blood vessel segment.

[0009] Furthermore, the process of obtaining the degree of abnormality in the changing trend includes: The number of pixels in the segmented region of each vessel segment when thresholding the coronary ultrasound image at each gray level threshold is taken as the number of segmented pixels of each vessel segment at each gray level threshold. The reference segmentation increment of each vessel segment at each gray level threshold is determined according to the difference between the number of segmented pixels of each vessel segment at each gray level threshold and the number of segmented pixels at the previous gray level threshold, in ascending order of gray level threshold. Based on the ratio between the reference segmentation increment and the total number of pixels in the corresponding blood vessel segment, the segmentation increment parameter of each blood vessel segment at each gray level threshold is determined; based on the standard deviation of the segmentation increment parameter of each blood vessel segment at all gray level thresholds between the corresponding first boundary threshold and the corresponding second boundary threshold, the change trend feature value of each blood vessel segment is determined. The overall trend characteristic value is determined based on the mean of the trend characteristic values ​​of all vascular segments; the difference between the trend characteristic value of each vascular segment and the overall trend characteristic value is normalized to determine the corresponding degree of trend abnormality.

[0010] Furthermore, the process of obtaining the probability of an anomaly includes: The probability of an anomaly in each blood vessel segment is determined based on the average of the degree of grayscale distribution abnormality, the degree of blood vessel diameter abnormality, and the degree of change trend abnormality.

[0011] Furthermore, the process of obtaining the abnormal vascular segment includes: The vascular segment with an abnormal probability greater than the preset abnormal threshold is defined as the abnormal vascular segment.

[0012] Furthermore, the process of obtaining the atherosclerotic plaque region includes: The abnormal vascular segment was segmented using the Otsu's method to determine the atherosclerotic plaque region.

[0013] Secondly, this application provides a system for segmenting atherosclerotic plaques in coronary ultrasound images, the system comprising: The image acquisition and preprocessing module is used to acquire coronary ultrasound images; and to perform edge segmentation on the coronary ultrasound images to determine all vessel segments. The first determining module is used to determine the segmentation threshold of each blood vessel segment based on the region display of each blood vessel segment during the threshold segmentation process of traversing the coronary ultrasound image; and to determine the degree of grayscale distribution abnormality of each blood vessel segment based on the segmentation threshold deviation between each blood vessel segment and adjacent blood vessel segments. The second determining module is used to determine the degree of abnormality of the blood vessel diameter of each blood vessel segment based on the deviation of the blood vessel width between each blood vessel segment and the adjacent blood vessel segments; and to determine the degree of abnormality of the change trend of each blood vessel segment based on the deviation of the change in the number of pixels of each blood vessel segment during the traversal threshold segmentation process. The atherosclerotic plaque region segmentation module is used to determine the probability of abnormality for each blood vessel segment based on the degree of abnormality in grayscale distribution, the degree of abnormality in blood vessel diameter, and the degree of abnormality in the trend of change; to screen out abnormal blood vessel segments based on the probability of abnormality; and to perform image segmentation on the abnormal blood vessel segments to determine the atherosclerotic plaque region.

[0014] Thirdly, this application provides a computer device including a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to perform the method as described in the first aspect of this application or any embodiment of the first aspect.

[0015] Fourthly, this application provides a computer program product comprising computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0016] Fifthly, this application provides a computer-readable storage medium that stores computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0017] This application has the following beneficial effects: After identifying the vascular segment, this application first determines the degree of grayscale distribution abnormality for each vascular segment based on the characteristic that adjacent normal vascular segments typically exhibit similar grayscale distributions. Further, based on the characteristic that the diameters of adjacent normal vascular segments usually do not change significantly, the application determines the degree of vascular diameter abnormality for each vascular segment. Then, based on the characteristic that the presence of atherosclerotic plaque regions disrupts the similarity of grayscale distribution in normal blood vessels, the application determines the degree of trend abnormality for each vascular segment. Finally, by combining the degree of grayscale distribution abnormality, the degree of vascular diameter abnormality, and the degree of trend abnormality, abnormal vascular segments affected by atherosclerotic plaques are screened out. Thus, image segmentation is performed only within abnormal vascular segments, resulting in more accurate segmentation of atherosclerotic plaque regions. Attached Figure Description

[0018] To more clearly illustrate the technical solutions and advantages 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.

[0019] Figure 1 A flowchart of a method for segmenting atherosclerotic plaques from coronary ultrasound images, provided in one embodiment of the present invention; Figure 2 An atherosclerotic plaque segmentation system for coronary ultrasound images is provided as an embodiment of the present invention; Figure 3 This is a schematic diagram of a computer device structure provided in one embodiment of the present invention. Detailed Implementation

[0020] 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 a method for segmenting atherosclerotic plaques in coronary ultrasound images according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment, and specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature.

[0021] 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.

[0022] The following describes in detail, with reference to the accompanying drawings, a specific scheme for the atherosclerotic plaque segmentation method in coronary ultrasound images provided by the present invention.

[0023] This application provides a method for segmenting atherosclerotic plaques in coronary ultrasound images. Please refer to [link to relevant documentation]. Figure 1 The diagram illustrates a flowchart of a method for segmenting atherosclerotic plaques from coronary ultrasound images according to an embodiment of the present invention. The method includes: Step S101: Acquire coronary ultrasound images; perform edge segmentation on the coronary ultrasound images to identify all vessel segments.

[0024] Initial coronary ultrasound images obtained through coronary ultrasound imaging technology are extracted from a medical database. After preprocessing the initial coronary ultrasound images, the coronary ultrasound images required in this embodiment of the invention are obtained. In a specific implementation of this embodiment, the process of acquiring coronary ultrasound images includes: performing median filtering on the initial coronary ultrasound images and then performing global histogram equalization to obtain the coronary ultrasound images. Median filtering can reduce speckle noise in the image, making the image clearer; global histogram equalization can improve contrast by stretching the image grayscale range, improving the visibility of blood vessel edges; and improving the accuracy of the blood vessel segments selected in subsequent analysis processes.

[0025] Furthermore, based on the characteristic that vascular regions in coronary ultrasound images typically correspond to a whole and have distinct edges, edge segmentation is performed on the coronary ultrasound images to determine all vascular segments required for subsequent analysis. Preferably, in some possible implementations of this invention, the process of obtaining vascular segments includes: Canny edge detection was performed on the coronary ultrasound image to determine the corresponding edge lines. The area enclosed by the edge lines with the largest area was taken as the vascular region. Morphological processing was performed on the vascular region to extract the skeleton and determine the corresponding vascular skeleton. On the vascular skeleton, the number of adjacent skeleton pixels of each skeleton pixel was taken as the corresponding reference number. Skeleton pixels with a reference number greater than 2 were used as segmentation points to divide the vascular skeleton into at least two skeleton segments. The corresponding vascular segment was determined based on the region corresponding to each skeleton segment in the vascular region.

[0026] Based on the principles of skeleton extraction and the structure of blood vessels, skeleton extraction simplifies the blood vessel region to a single-pixel-wide centerline, facilitating subsequent vessel segmentation. For each skeleton pixel on the blood vessel skeleton, a reference number greater than 2 usually indicates the presence of a blood vessel branch at that pixel location. Therefore, skeleton pixels with a reference number greater than 2 are used as segmentation points, making the blood vessel segments determined based on skeleton segmentation more accurate. For each skeleton segment, the region of each segment before morphological processing is taken as the corresponding blood vessel segment. It should be noted that Canny edge detection and morphological processing for skeleton extraction are image processing techniques well-known to those skilled in the art and will not be elaborated further here.

[0027] Step S102: Determine the segmentation threshold for each blood vessel segment based on the region display of each segment during the threshold segmentation process of traversing the coronary ultrasound image; determine the degree of grayscale distribution abnormality of each blood vessel segment based on the segmentation threshold deviation between each blood vessel segment and its adjacent segments.

[0028] In coronary ultrasound images, blood vessels gradually narrow along the direction of blood flow, resulting in a decrease in blood volume. This causes the grayscale value of the vascular region in the image to gradually increase, and the change in grayscale value is usually small. Therefore, when adjacent normal vascular segments are segmented using thresholding, the grayscale segmentation thresholds that can be used to segment the corresponding vascular segments are usually similar. Therefore, after further determining the segmentation threshold based on the regional manifestation of each vascular segment during the thresholding process of coronary ultrasound images, the greater the deviation in segmentation threshold between the corresponding vascular segment and adjacent vascular segments, the more likely there is an abnormal grayscale distribution. The corresponding segment is less likely to conform to the characteristics of a normal vascular segment and is more likely to be affected by atherosclerotic plaques, that is, it is more likely to be an abnormal vascular segment.

[0029] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the segmentation threshold includes: The minimum gray value of the coronary ultrasound image is used as the starting gray value threshold; the maximum gray value of the coronary ultrasound image is used as the ending gray value threshold; all gray values ​​between the starting gray value threshold and the ending gray value threshold are traversed using a preset gray value interval as the traversal step size to determine all gray value thresholds; in a specific implementation of this invention, the preset gray value interval is set to 1, which can be adjusted according to the specific implementation environment.

[0030] In the process of threshold segmentation of coronary ultrasound images in ascending order of grayscale thresholds, the grayscale threshold of the first segmented region in each vessel segment is used as the first boundary threshold for each vessel segment; the grayscale threshold of the first complete segmentation of each vessel segment is used as the second boundary threshold for each vessel segment; and the segmentation threshold for each vessel segment is determined based on the average of the first and second boundary thresholds. The first appearance of a segmented region means that the corresponding grayscale threshold exactly matches the bottom grayscale feature of the corresponding vessel segment, while the first complete segmentation of the corresponding vessel segment indicates that the corresponding grayscale threshold exactly matches the highest grayscale feature of the corresponding vessel segment. The bottom and highest grayscale features of adjacent normal vessel segments are usually close to consistent, while the high grayscale characteristics exhibited in atherosclerotic plaque regions usually destroy the highest grayscale feature, resulting in a significant difference between the segmentation threshold of abnormal vessel segments with atherosclerotic plaque regions and the segmentation thresholds of other adjacent vessel segments; therefore, the degree of grayscale distribution abnormality of each vessel segment can be further determined based on the segmentation threshold deviation between each vessel segment and its adjacent vessel segments.

[0031] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the degree of grayscale distribution anomaly includes: Each vascular segment is sequentially designated as the target segment; all adjacent vascular segments are designated as the corresponding reference segments; based on the difference between the segmentation threshold of the target segment and the segmentation threshold of each reference segment, the threshold deviation value of each reference segment is determined; the larger the threshold deviation value, the less the segmentation threshold of the target segment conforms to the characteristic that the segmentation thresholds between adjacent normal vascular segments are usually quite close compared to the corresponding reference segment, indicating that the target segment is more likely to exhibit abnormal grayscale distribution under the comparison with the corresponding reference segment.

[0032] Considering that abnormal vascular segments with atherosclerotic plaques may also appear in the various reference segments of the target segment, and that the segmentation thresholds between adjacent normal vascular segments are usually quite similar, for each reference segment, the larger the overall deviation of its segmentation threshold from all other reference segments, the more likely the corresponding reference segment is to be abnormal, and the lower its reference value. Therefore, the mean of the segmentation threshold differences between each reference segment and all other reference segments is negatively correlated and normalized to determine the reference weight of each reference segment. The reference weight is used to measure the reference value of the threshold deviation value of each reference segment. After determining the weighted deviation value of each reference segment by multiplying the reference weight and the threshold deviation value, the resulting weighted deviation value can more accurately characterize the gray-scale distribution anomaly of the target segment and reduce the impact of abnormal reference segments on the calculation results. Finally, the mean of the weighted deviation values ​​of all reference segments corresponding to the target segment is normalized by combining the weighted deviation values ​​of all reference segments to determine the degree of gray-scale distribution anomaly of the target segment.

[0033] In one specific implementation of this invention, the process of obtaining the reference weights includes: ;in, For reference Reference weights; For reference Of all the reference segments corresponding to the target segment, excluding the reference segment The number of other reference segments besides; For reference The segmentation threshold; For reference Of all the reference segments corresponding to the target segment, excluding the reference segment The outside The segmentation threshold for each reference segment; It is the absolute value symbol; It is a linear normalization function; for The function aims to make the sum of the reference weights of all reference segments corresponding to the target segment equal to 1. It should be noted that, unless otherwise specified, the normalization method in the embodiments of this invention uses linear normalization, which will not be further elaborated here.

[0034] In one specific implementation of this invention, the process of obtaining the degree of grayscale distribution anomaly is expressed by the formula: ;in, For the target segment The degree of abnormality in grayscale distribution; For the target segment The number of reference segments; For the target segment The segmentation threshold; For the target segment The corresponding number The segmentation threshold for each reference segment; It is the absolute value symbol; For the target segment The corresponding number Threshold deviation values ​​for each reference segment; For the target segment The corresponding number The reference weight of each reference segment; For the target segment The corresponding number The weighted deviation value of each reference segment.

[0035] Step S103: Determine the degree of abnormality of the blood vessel diameter of each blood vessel segment based on the deviation of the blood vessel width between each blood vessel segment and its adjacent blood vessel segments; determine the degree of abnormality of the change trend of each blood vessel segment based on the deviation of the change in the number of pixels in each blood vessel segment during the threshold segmentation process.

[0036] Considering that the diameter of a normal blood vessel segment typically changes relatively smoothly during its extension, meaning that the diameters of adjacent normal blood vessel segments are usually quite similar; however, the presence of atherosclerotic plaques disrupts the width characteristics of the segment. Therefore, for each blood vessel segment, the greater the deviation in diameter between its segment and that of its adjacent segments, the less the segment conforms to the characteristics of a normal blood vessel segment, and the higher the likelihood of it being affected by atherosclerotic plaques. Therefore, the degree of abnormality in the blood vessel diameter of each segment is further determined based on the deviation in blood vessel width between each segment and its adjacent segments. A greater degree of abnormality in the blood vessel diameter indicates that the corresponding segment is more likely to be an abnormal blood vessel segment.

[0037] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the degree of abnormality in blood vessel diameter includes: The diameter of a blood vessel is determined based on the overall relative distance between the two edge lines on both sides of the vascular skeleton of each segment. The process of obtaining the blood vessel diameter includes: in each segment, the two edge lines on both sides of the vascular skeleton are taken as the vessel edge lines; the minimum distance between each pixel on each vessel edge line and all pixels on the other vessel edge line is taken as the local distance of each pixel on each vessel edge line; the diameter of each segment is determined based on the average of the local distances of all pixels on the two vessel edge lines of each segment. The local distance represents the shortest distance required for each pixel on the vessel edge line to cross the segment, and it is usually perpendicular to the direction of vessel extension, thus representing the local width characteristics of the vessel. Considering that the diameter or width characteristics of a blood vessel segment differ at different locations, the blood vessel diameter is determined using the overall local distance characteristics of each edge pixel of the segment.

[0038] Based on the average diameter of all adjacent vascular segments, a reference diameter for each vascular segment is determined. Since the diameters of adjacent normal vascular segments are usually similar, the greater the deviation between the diameter of a vascular segment and its corresponding reference diameter, the more abnormal the vascular segment may be. Therefore, the difference between the diameter of each vascular segment and the reference diameter is further normalized to determine the degree of abnormality of the diameter of each vascular segment.

[0039] Preferably, in a specific implementation of this invention, the process of obtaining the degree of abnormality in blood vessel diameter includes: ;in, For vascular segment The degree of abnormality in blood vessel diameter; For vascular segment The diameter of the blood vessels; For vascular segment The number of all adjacent vascular segments; For vascular segment The adjacent first The diameter of the blood vessel segment; For vascular segment The reference diameter.

[0040] Furthermore, considering that the structure of normal blood vessel segments is usually relatively regular, during the threshold segmentation process, the proportion of pixel changes across all segmentation thresholds in normal blood vessel segments typically exhibits a similar distribution. However, the grayscale distribution pattern within atherosclerotic plaque regions is usually different from that of normal blood vessels. This means that when atherosclerotic plaques are present in a blood vessel segment, the proportion of pixel changes can abruptly change with the segmentation threshold, thereby disrupting the distribution characteristics of pixel changes corresponding to normal blood vessel segments. Therefore, based on the deviation in the number of pixels for each blood vessel segment during the threshold segmentation process, the degree of abnormality in the trend of change for each segment is determined. The greater the degree of abnormality in the trend of change, the higher the influence of atherosclerotic plaques.

[0041] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the degree of abnormality in the trend of change includes: The number of pixels in the segmented region of each vessel segment when performing threshold segmentation on the coronary ultrasound image at each grayscale threshold is taken as the number of segmented pixels for each vessel segment at each grayscale threshold. Following the order of grayscale thresholds from smallest to largest, the reference segmentation increment for each vessel segment at each grayscale threshold is determined based on the difference between the number of segmented pixels for each vessel segment at each grayscale threshold and the number of segmented pixels at the previous grayscale threshold. The segmentation increment parameter for each vessel segment at each grayscale threshold is determined based on the ratio between the reference segmentation increment and the total number of pixels in the corresponding vessel segment. The trend characteristic value of each vessel segment is determined based on the standard deviation of the segmentation increment parameter for each vessel segment at all grayscale thresholds between the corresponding first boundary threshold and the corresponding second boundary threshold.

[0042] Normalizing the reference segmentation increment by the total number of pixels in each blood vessel segment can reduce the impact of blood vessel segments with different numbers of pixels on the calculation process of segmentation increment parameters. The standard deviation can characterize the distribution characteristics of a set of data in the data distribution dimension. Therefore, the trend feature value characterizes the distribution characteristics of the segmentation increment parameters of each blood vessel segment, that is, the distribution characteristics of the pixel change ratio at all segmentation thresholds.

[0043] Considering that the pixel change ratio distribution characteristics are usually similar among normal blood vessel segments, while the pixel change ratio distribution characteristics of blood vessel segments with atherosclerotic plaques will deviate from those of normal blood vessel segments, and considering that blood vessel segments with atherosclerotic plaques are usually fewer, an overall trend feature value is determined based on the mean of the trend feature values ​​of all blood vessel segments. This overall trend feature value is used to characterize the pixel change ratio distribution characteristics of normal blood vessel segments. Finally, the difference between the trend feature value of each blood vessel segment and the overall trend feature value is normalized to determine the corresponding degree of trend abnormality. The greater the degree of trend abnormality, the more the pixel change ratio distribution characteristics of the corresponding blood vessel segment deviate from those of normal blood vessel segments, meaning that the corresponding blood vessel segment is more affected by atherosclerotic plaques and is more likely to be an abnormal blood vessel segment.

[0044] In one specific implementation of this invention, the process of obtaining the degree of abnormality in the trend of change is expressed by the following formula: ;in, For vascular segment The degree of abnormality in the trend of change; For vascular segment The standard deviation of the segmentation increment parameter at all grayscale thresholds between the corresponding first boundary threshold and the corresponding second boundary threshold, i.e., the blood vessel segment The characteristic value of the changing trend; This is the mean of the trend characteristic values ​​of all blood vessel segments, which is also the overall trend characteristic value.

[0045] Step S104: Determine the probability of abnormality for each blood vessel segment based on the degree of abnormality in grayscale distribution, blood vessel diameter, and trend of change; screen out abnormal blood vessel segments based on the probability of abnormality; perform image segmentation on the abnormal blood vessel segments to determine the atherosclerotic plaque region.

[0046] Since the degree of abnormality in grayscale distribution, the degree of abnormality in vessel diameter, and the degree of abnormality in change trend can all characterize the likelihood that a vessel segment belongs to an abnormal vessel segment, we further combine these three parameters to determine the probability of abnormality for each vessel segment, thereby screening out more accurate abnormal vessel segments affected by atherosclerotic plaques based on the probability of abnormality.

[0047] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the probability of an anomaly includes: The probability of abnormality for each blood vessel segment is determined by the average of the abnormality in grayscale distribution, the abnormality in blood vessel diameter, and the abnormality in change trend.

[0048] Since the abnormality levels of grayscale distribution, vessel diameter, and trend of change in this embodiment are all normalized values, the probability of abnormality is determined by the mean, limiting the range of the probability of abnormality to 0 to 1, which facilitates the subsequent screening of abnormal vessel segments. In another specific implementation of this embodiment, the product of the abnormality levels of grayscale distribution, vessel diameter, and trend of change is normalized to determine the probability of abnormality for each vessel segment; that is, after fusing the three parameters of abnormality levels of grayscale distribution, vessel diameter, and trend of change through product, the corresponding value results are corrected by normalization, which will not be elaborated further here.

[0049] In one specific implementation of this invention, the process of obtaining the probability of an anomaly is expressed by the following formula: ;in, For vascular segment The degree of abnormality in grayscale distribution; For vascular segment The degree of abnormality in blood vessel diameter; For vascular segment The degree of abnormality in the trend of change.

[0050] In one specific implementation of this invention, the process of obtaining abnormal vascular segments includes: identifying vascular segments with an abnormal probability greater than a preset abnormal threshold as abnormal vascular segments, i.e., vascular segments containing atherosclerotic plaque regions. The preset abnormal threshold is set to 0.5 and can be adjusted according to the specific implementation environment.

[0051] After identifying the abnormal vascular segment containing atherosclerotic plaques, the abnormal vascular segment can be segmented based on the relatively high grayscale value of the atherosclerotic plaque region to determine the atherosclerotic plaque region. Preferably, in some possible implementations of this invention, the process of obtaining the atherosclerotic plaque region includes: segmenting the abnormal vascular segment using the maximum inter-class variance (MOV) method to determine the atherosclerotic plaque region; specifically, after determining the OTSU threshold of the corresponding abnormal vascular segment using the MOV method, the connected region composed of pixels with grayscale values ​​greater than the OTSU threshold in the abnormal vascular segment is taken as the atherosclerotic plaque region. It should be noted that the MOV method is a well-known technique among those skilled in the art and will not be further elaborated here.

[0052] In summary, a method for segmenting atherosclerotic plaques in coronary ultrasound images, after identifying the vessel segment, first determines the degree of grayscale distribution abnormality in each vessel segment based on the characteristic that adjacent normal vessel segments typically exhibit similar grayscale distributions. Further, based on the characteristic that the diameters of adjacent normal vessel segments usually do not show significant changes, the method determines the degree of vessel diameter abnormality in each vessel segment. Then, based on the characteristic that the presence of atherosclerotic plaque regions disrupts the similarity of grayscale distribution in normal vessels, the method determines the degree of trend abnormality in each vessel segment. Finally, by combining the degree of grayscale distribution abnormality, the degree of vessel diameter abnormality, and the degree of trend abnormality, abnormal vessel segments affected by atherosclerotic plaques are screened out. Thus, image segmentation is performed only in abnormal vessel segments, resulting in more accurate segmentation of atherosclerotic plaque regions.

[0053] This application also provides a system for segmenting atherosclerotic plaques in coronary ultrasound images. Please refer to [link to relevant documentation]. Figure 2 The diagram shows a structural diagram of an atherosclerotic plaque segmentation system for coronary ultrasound images provided in an embodiment of the present invention. The system includes: an image acquisition and preprocessing module 201, a first determination module 202, a second determination module 203, and an atherosclerotic plaque region segmentation module 204.

[0054] The image acquisition and preprocessing module 201 is used to acquire coronary ultrasound images; perform edge segmentation on the coronary ultrasound images to determine all vessel segments; The first determining module 202 is used to determine the segmentation threshold of each blood vessel segment based on the region display of each blood vessel segment during the threshold segmentation process of traversing the coronary ultrasound image; and to determine the degree of grayscale distribution abnormality of each blood vessel segment based on the segmentation threshold deviation between each blood vessel segment and adjacent blood vessel segments. The second determining module 203 is used to determine the degree of abnormality of the blood vessel diameter of each blood vessel segment based on the deviation of the blood vessel width between each blood vessel segment and the adjacent blood vessel segments; and to determine the degree of abnormality of the change trend of each blood vessel segment based on the deviation of the change in the number of pixels of each blood vessel segment during the traversal threshold segmentation process. The atherosclerotic plaque region segmentation module 204 is used to determine the probability of abnormality of each blood vessel segment based on the degree of abnormality in grayscale distribution, the degree of abnormality in blood vessel diameter, and the degree of abnormality in change trend; to screen out abnormal blood vessel segments based on the probability of abnormality; and to perform image segmentation on the abnormal blood vessel segments to determine the atherosclerotic plaque region.

[0055] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the atherosclerotic plaque segmentation system for coronary ultrasound images and the atherosclerotic plaque segmentation method for coronary ultrasound images provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0056] This application also provides a computer device; please refer to [link / reference]. Figure 3 The illustration shows a schematic diagram of a computer device structure according to an embodiment of the present invention. The computer device includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any of the coronary ultrasound image atherosclerotic plaque segmentation methods described above.

[0057] This application also provides a computer program product that, when run on a computer device, enables the computer device to execute any of the aforementioned methods for segmenting atherosclerotic plaques from coronary ultrasound images.

[0058] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer device, the computer device can execute any of the aforementioned methods for segmenting atherosclerotic plaques from coronary ultrasound images.

[0059] In the embodiments provided in this application, it should be understood that the computer device, computer program product and computer-readable storage medium provided are all used to perform the corresponding methods provided above, and therefore the beneficial effects they can achieve can be referred to the beneficial effects of the methods provided above, which will not be repeated here.

[0060] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0061] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for segmenting atherosclerotic plaques from coronary ultrasound images, characterized in that, The method includes: Acquire coronary ultrasound images; perform edge segmentation on the coronary ultrasound images to identify all vessel segments; Based on the region display of each vessel segment during the threshold segmentation process of the coronary ultrasound image, the segmentation threshold of each vessel segment is determined; based on the segmentation threshold deviation between each vessel segment and its adjacent segments, the degree of grayscale distribution abnormality of each vessel segment is determined. The degree of abnormality in the diameter of each blood vessel segment is determined based on the deviation in the width of each blood vessel segment from its adjacent segments; the degree of abnormality in the trend of change of each blood vessel segment is determined based on the deviation in the number of pixels of each blood vessel segment during the threshold segmentation process. Based on the degree of abnormality in grayscale distribution, the degree of abnormality in vessel diameter, and the degree of abnormality in the trend of change, the probability of abnormality for each vessel segment is determined; abnormal vessel segments are screened out based on the probability of abnormality; image segmentation is performed on the abnormal vessel segments to determine the atherosclerotic plaque region.

2. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the blood vessel segment includes: The coronary ultrasound image is subjected to Canny edge detection to determine the corresponding edge lines; the area enclosed by the edge lines with the largest area is taken as the vascular region; the vascular region is subjected to morphological processing to extract the skeleton and determine the corresponding vascular skeleton. On the vascular skeleton, the number of adjacent skeleton pixels of each skeleton pixel is used as the corresponding reference number; the vascular skeleton is divided into at least two skeleton segments using skeleton pixels with a reference number greater than 2 as segmentation points; and the corresponding vascular segment is determined according to the region corresponding to each skeleton segment in the vascular region.

3. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the segmentation threshold includes: The minimum gray value of the coronary ultrasound image is used as the starting gray value threshold; the maximum gray value of the coronary ultrasound image is used as the ending gray value threshold; all gray values ​​between the starting gray value threshold and the ending gray value threshold are traversed with a preset gray value interval as the traversal step size to determine all gray value thresholds. In the process of traversing and segmenting coronary ultrasound images in ascending order of grayscale threshold, the grayscale threshold of the first segmented region in each vessel segment is used as the first boundary threshold of each vessel segment; the grayscale threshold of the first complete segmentation of each vessel segment is used as the second boundary threshold of each vessel segment; and the segmentation threshold of each vessel segment is determined based on the average value between the first boundary threshold and the second boundary threshold.

4. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the degree of grayscale distribution anomaly includes: Each vascular segment is sequentially designated as the target segment; all vascular segments adjacent to the target segment are designated as the corresponding reference segments; the threshold deviation value of each reference segment is determined based on the difference between the segmentation threshold of the target segment and the segmentation threshold of each reference segment; the mean of the segmentation threshold differences between each reference segment and all other reference segments is negatively correlated and normalized to determine the reference weight of each reference segment. The weighted deviation value of each reference segment is determined by multiplying the reference weight by the threshold deviation value; the mean of the weighted deviation values ​​of all reference segments corresponding to the target segment is normalized to determine the degree of grayscale distribution anomaly of the target segment.

5. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 2, characterized in that, The process of obtaining the degree of abnormality in the blood vessel diameter includes: The corresponding blood vessel diameter is determined based on the overall relative distance between the two edge lines on both sides of the vascular skeleton of each blood vessel segment; the reference diameter of each blood vessel segment is determined based on the average blood vessel diameter of all adjacent blood vessel segments; the difference between the blood vessel diameter of each blood vessel segment and the reference diameter is normalized to determine the degree of abnormality of the blood vessel diameter of each blood vessel segment.

6. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 5, characterized in that, The process of obtaining the blood vessel diameter includes: In each blood vessel segment, the two edge lines on both sides of the blood vessel skeleton are taken as the blood vessel edge lines; the minimum distance between each pixel on each blood vessel edge line and all pixels on the other blood vessel edge line is taken as the local distance of each pixel on each blood vessel edge line; the blood vessel diameter of each blood vessel segment is determined based on the average of the local distances of all pixels on the two blood vessel edge lines of each blood vessel segment.

7. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 3, characterized in that, The process of obtaining the degree of abnormality in the trend of change includes: The number of pixels in the segmented region of each vessel segment when thresholding the coronary ultrasound image at each gray level threshold is taken as the number of segmented pixels of each vessel segment at each gray level threshold. The reference segmentation increment of each vessel segment at each gray level threshold is determined according to the difference between the number of segmented pixels of each vessel segment at each gray level threshold and the number of segmented pixels at the previous gray level threshold, in ascending order of gray level threshold. Based on the ratio between the reference segmentation increment and the total number of pixels in the corresponding blood vessel segment, the segmentation increment parameter of each blood vessel segment at each gray level threshold is determined; based on the standard deviation of the segmentation increment parameter of each blood vessel segment at all gray level thresholds between the corresponding first boundary threshold and the corresponding second boundary threshold, the change trend feature value of each blood vessel segment is determined. The overall trend characteristic value is determined based on the mean of the trend characteristic values ​​of all vascular segments; the difference between the trend characteristic value of each vascular segment and the overall trend characteristic value is normalized to determine the corresponding degree of trend abnormality.

8. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the probability of an anomaly includes: The probability of an anomaly in each blood vessel segment is determined based on the average of the degree of grayscale distribution abnormality, the degree of blood vessel diameter abnormality, and the degree of change trend abnormality.

9. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the abnormal vascular segment includes: The vascular segment with an abnormal probability greater than the preset abnormal threshold is defined as the abnormal vascular segment.

10. The method for segmenting atherosclerotic plaques in coronary ultrasound images according to claim 1, characterized in that, The process of obtaining the atherosclerotic plaque region includes: The abnormal vascular segment was segmented using the Otsu's method to determine the atherosclerotic plaque region.

Images