A carotid artery plaque echo detection method, device, medium and product

Abstract

The present invention discloses a method, device, medium and product for detecting echo of carotid plaque, and relates to the field of medical image detection. The method comprises: dividing the region of interest of the carotid plaque ultrasound image into multiple sub-regions; calculating the statistics of each sub-region; the statistics include: distribution, skewness and kurtosis; constructing the sub-region vector of each sub-region according to the statistics; clustering the sub-region vectors of all sub-regions using the improved k-means algorithm to obtain clustering results; the improved k-means algorithm is a k-means algorithm that adds weights when calculating the distance between the sub-region vector and the center vector; the center vector is randomly generated; the clustering result includes k clusters, each cluster includes sub-region vectors with similar echoes; and the echo detection results of each sub-region are determined according to the area of ​​each cluster in the clustering result. The present invention can accurately and efficiently realize the detection of carotid plaque echo.

Description

Technical Field

[0001] The present invention relates to the field of medical image detection, and in particular to a method, device, medium and product for detecting carotid artery plaque echo. Background Art

[0002] Carotid plaques are manifestations of carotid atherosclerosis. They are clumps of cholesterol, fat, and blood cells that often occur at the bifurcation of the common carotid artery and are believed to be closely related to the occurrence of ischemic stroke in the elderly. Carotid artery disease occurs when plaques block the carotid arteries that carry blood to the brain and head. Blockages increase the risk of stroke. Stroke is an emergency that occurs when the brain loses all or most of its blood supply. During a stroke, the brain does not get oxygen and brain cells begin to die in just a few minutes. The formation of carotid plaques is affected by many factors, among which age > 60 years old, male sex, long-term smoking history, history of hypertension, history of diabetes, and hyperlipidemia are the main risk factors for the formation of carotid plaques. Therefore, it is particularly important to detect the echo of carotid plaques.

[0003] The commonly used method for detecting carotid plaque echoes is ultrasonic detection. Ultrasound is used to scan the patient's carotid artery. Then the doctor manually judges the ultrasound image with the naked eye to identify whether there is a plaque in the image, and judges the characteristics of the plaque, and finally describes the characteristics in the diagnosis report. However, this approach has the following problems: 1) The diagnostic efficiency is slow, and the doctor needs to scan the patient's affected area while identifying and interpreting; 2) When writing a diagnostic report, it is time-consuming and laborious; 3) The doctor's workload is large, and he needs to concentrate on each patient, which consumes a lot of physical energy; 4) The doctor's experience level is high, and the training cycle of professional ultrasound doctors is long and the cost is high; 5) The number of competent doctors is limited, medical resources are tight and the cost is high. It can be seen that the above-mentioned method of manually judging plaque echoes requires rich clinical experience, strong concentration and memory, which is time-consuming, labor-intensive and costly. In order to overcome the above problems, it is proposed to detect carotid plaque echoes based on a supervised learning algorithm, but this method requires a lot of time and cost to collect data, and the error is affected by the sample.

[0004] In summary, how to accurately and efficiently detect carotid artery plaque echoes has become a problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of the present invention is to provide a method, device, medium and product for detecting carotid artery plaque echo, which can accurately and efficiently detect carotid artery plaque echo.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for detecting carotid artery plaque echo, the method comprising:

[0008] Obtain ultrasound images of carotid artery plaques;

[0009] Segmenting the region of interest of the carotid plaque ultrasound image into a plurality of sub-regions;

[0010] Calculating statistics of each of the sub-regions; the statistics include: distribution, skewness and kurtosis;

[0011] constructing a sub-region vector for each of the sub-regions according to the statistics;

[0012] The sub-region vectors of all sub-regions are clustered using an improved k-means algorithm to obtain a clustering result; the improved k-means algorithm is a k-means algorithm that adds weights when calculating the distance between the sub-region vector and the center vector; the center vector is randomly generated; the clustering result includes k clusters, each of which includes sub-region vectors with similar echoes;

[0013] The echo detection result of each sub-region is determined according to the area of ​​each cluster in the clustering result.

[0014] Optionally, the sub-region vectors of all sub-regions are clustered using an improved k-means algorithm to obtain clustering results, specifically including:

[0015] Randomly generate k center vectors with different k values; where k = 3, 4, 5, ...;

[0016] For any value of k, the distance between the matrix elements of each sub-region vector and the corresponding matrix elements of the k center vectors is calculated, and a weight is assigned to the distance between each matrix element of the sub-region vector and the corresponding matrix element of the k center vectors to obtain the final distance between the sub-region vector and the k center vectors; each parameter in the statistic is taken as a matrix element;

[0017] For any k value, according to the final distance, with the goal of minimizing the partial square sum of errors, the values ​​of the center vector and the weight are optimized cyclically and alternately to obtain the optimal values ​​of the weight and k center vectors;

[0018] The value of k is changed, and the optimal value of k is determined by the elbow point method; the optimal value is the k value corresponding to the inflection point in the k-SSE line graph; the k-SSE line graph represents the relationship between different k values ​​and the minimum partial sum of squares of error;

[0019] When determining the optimal value of k, the final distance between the sub-region vector and the k center vectors is calculated using the weight and the optimal values ​​of the k center vectors;

[0020] The echo category of each sub-region vector is determined according to the final distance between all sub-region vectors and the k center vectors when k takes the optimal value, and the sub-region vectors with the same type of echo are determined as a cluster in the clustering result.

[0021] Optionally, determining the echo detection result of each sub-region according to the area of ​​each cluster in the clustering result specifically includes:

[0022] If k=3, the three clusters are optimized to obtain optimized clusters, and the optimized clusters are used as the echo detection results of each of the sub-regions;

[0023] If k≥4, the cluster with the smallest area is merged into the cluster closest to the cluster until k=3, and then the three clusters are optimized to obtain an optimized cluster, and the optimized cluster is used as the echo detection result of each sub-region;

[0024] The process of optimizing the three clusters includes:

[0025] The three clusters are identified as a low-echo cluster, an isoechoic cluster, and a high-echo cluster in the order of average pixel values ​​from low to high;

[0026] The clusters whose areas are lower than a set proportion of the area of ​​the region of interest among the low-echo cluster, the isoechoic cluster and the high-echo cluster are merged into the cluster closest to the cluster to obtain an optimized cluster.

[0027] Optionally, if the matrix element is a distribution, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Wasserstein distance; if the matrix element is a skewness or a kurtosis, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Euclidean distance.

[0028] Optionally, the region of interest of the carotid artery plaque ultrasound image is divided into a plurality of sub-regions, specifically including:

[0029] The region of interest of the carotid artery plaque ultrasound image is segmented into multiple sub-regions using a superpixel algorithm.

[0030] Optionally, the set ratio is 20%.

[0031] The present invention also provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-mentioned carotid artery plaque echo detection method.

[0032] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned carotid artery plaque echo detection method are implemented.

[0033] The present invention also provides a computer program product, comprising a computer program, which implements the steps of the above-mentioned carotid artery plaque echo detection method when executed by a processor.

[0034] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0035] The embodiment of the present invention constructs the sub-region vector of each sub-region by calculating the distribution, skewness and kurtosis of each sub-region, and adopts the improved k-means algorithm to realize the detection of carotid plaque echo. The improved k-means algorithm is a k-means algorithm that adds weights when calculating the distance between the sub-region vector and the center vector. The vector space is constructed by using distribution, skewness and kurtosis, which greatly reduces the influence of sample discrete values, abnormal values ​​and extreme values ​​compared with the traditional commonly used brightness, contrast, etc., and improves the accuracy of detection; based on unsupervised learning, it is not affected by the global parameters of the image (such as overall brightness, contrast, resolution), has strong generalization, and does not need to collect training data, avoiding the waste of time and cost. The improved k-means algorithm is adopted, which has higher accuracy and stronger robustness than the traditional k-means algorithm, and further improves the accuracy of detection. Therefore, the embodiment of the present invention can accurately and efficiently realize the detection of carotid plaque echo. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A schematic diagram of the process of the carotid artery plaque echo detection method provided in Example 1 of the present invention;

[0038] Figure 2 The ultrasonic image and ROI schematic diagram of the carotid artery plaque provided in Example 1 of the present invention;

[0039] Figure 3 A schematic diagram of multiple sub-regions of ROI provided in Embodiment 1 of the present invention;

[0040] Figure 4 The k-SSE line graph provided in Example 1 of the present invention;

[0041] Figure 5 A schematic diagram of a cluster optimization process provided in Example 1 of the present invention;

[0042] Figure 6A specific implementation process diagram of a method for detecting carotid artery plaque echo;

[0043] Figure 7 This is a diagram of the internal structure of a computer device. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The purpose of the present invention is to provide a method, device, medium and product for detecting carotid plaque echo, aiming to accurately and efficiently realize the detection of carotid plaque echo.

[0046] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] Example 1

[0048] like Figure 1 As shown, the carotid artery plaque echo detection method in this embodiment includes:

[0049] Step 101: Acquire a carotid artery plaque ultrasound image.

[0050] Step 102: Segment the region of interest (ROI) of the carotid artery plaque ultrasound image into multiple sub-regions.

[0051] Specifically, a superpixel algorithm is used to segment the region of interest of the carotid artery plaque ultrasound image into multiple sub-regions. Figure 2 The ultrasound image of the carotid artery plaque is shown. The region of interest (ROI) of the ultrasound image of the arterial plaque is determined according to the outline coordinate points of the plaque in the figure. The plaque area within the outline points is the ROI. Figure 3 A plurality of sub-regions formed after segmentation are shown.

[0052] Step 103: Calculate the statistics of each of the sub-regions; the statistics include: distribution, skewness and kurtosis.

[0053] The calculation formula of the statistic is as follows:

[0054]

[0055] Among them, x represents the pixel value in the sub-region; σ represents the sample standard deviation; μ represents the sample mean; and F(x) represents the distribution of pixel values ​​in the sub-region.

[0056]

[0057] Where n represents the sample size; S(x) represents the skewness of the pixel values ​​in the sub-region.

[0058]

[0059] K(x) represents the kurtosis of the pixel values ​​in the sub-region.

[0060] Step 104: constructing a sub-region vector for each of the sub-regions according to the statistics.

[0061] Specifically, a vector space is constructed by distribution F, skewness S, and kurtosis K. The distribution, skewness, and kurtosis of each sub-region are expressed as the sub-region vector A of the vector space:

[0062]

[0063] Step 105: clustering the sub-region vectors of all sub-regions using an improved k-means algorithm to obtain clustering results.

[0064] The improved k-means algorithm is a k-means algorithm that adds weights when calculating the distance between the sub-region vector and the center vector; the center vector is randomly generated; and the clustering result includes k clusters, each of which includes sub-region vectors with similar echoes.

[0065] Step 105 specifically includes:

[0066] ① Randomly generate k center vectors C1, C2...C with different k values k ; Wherein, k = 3, 4, 5... Because the echo results are divided into 3 categories, k starts from 3.

[0067] ② For any value of k, calculate the distance between the matrix elements of each sub-region vector and the corresponding matrix elements of the k center vectors, and assign a weight to the distance between each matrix element of the sub-region vector and the corresponding matrix element of the k center vectors to obtain the final distance between the sub-region vector and the k center vectors. Each parameter in the statistic is taken as a matrix element.

[0068] Specifically, if the matrix element is distributed, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Wasserstein distance, and the calculation formula is:

[0069]

[0070] Among them, d F [p,q] represents the Wasserstein distance between p and q, where p and q represent two distributions whose distances need to be calculated; Π[p,q] represents the joint distribution of p and q; inf represents taking the lower bound; γ(x,y) represents an element in the joint distribution Π[p,q]; the above formula is for double integral, d(x,y) is the sign of the first integral, and dxdy is the sign of the second integral.

[0071] If the matrix element is skewness or kurtosis, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Euclidean distance, calculated as:

[0072]

[0073]

[0074] Among them, d S represents the Euclidean distance between the skewness S1 of the sub-region vector and the skewness S2 corresponding to the center vector; d K It represents the Euclidean distance between the kurtosis K1 of the sub-region vector and the kurtosis K2 corresponding to the center vector.

[0075] The distance matrix D formed by the above calculation results is:

[0076]

[0077] Since the distribution F, skewness S and kurtosis K are statistics of different orders and have different significance for clustering, the distance between the sub-region vector and the center vector cannot be a simple addition of the elements of the distance matrix, but a weight W must be added:

[0078]

[0079] The final distance between the sub-region vector and the center vector is:

[0080] distance=W T D.

[0081] Among them, w F 、w S 、w K They correspond to the weights of distribution, skewness, and kurtosis respectively; T means transpose.

[0082] ③For any value of k, according to the final distance, with the goal of minimizing the sum of squared errors (SSE), the values ​​of the center vector and weight are optimized cyclically and alternately to obtain the optimal values ​​of the weight and k center vectors.

[0083] Compared with the traditional k-means algorithm, the improved k-means algorithm in this embodiment needs to iteratively optimize not only the value of the center vector C, but also the value of the weight W. Specifically: the center vector C and the weight W are optimized alternately in a loop until the center vector C and the weight W no longer change, that is, converge to the optimal value. The optimization goal is to minimize the sum of squared errors, and the objective function is as follows:

[0084]

[0085] Where i represents the number of the center vector; C i represents the set of all sub-region vectors in the cluster represented by the i-th center vector; A represents the sub-region vector; distance(A,C i ) represents A and C i The final distance.

[0086] ④ Change the value of k and use the elbow point method to determine the optimal value of k; the optimal value is the k value corresponding to the inflection point in the k-SSE line graph; the k-SSE line graph represents the relationship between different k values ​​and the minimum partial sum of squares of errors.

[0087] Specifically, the above operation is repeated under different k values ​​to obtain the minimum SSE for different k values, and the optimal value of k is determined by the elbow point method. Figure 4 It can be seen from the k-SSE line graph shown that when k<4, the curve drops rapidly; when k>4, the curve tends to be stable, and the inflection point 4 is determined to be the optimal value of k through the elbow point method.

[0088] That is, the difference in slope before and after the maximum k point is the largest, expressed as:

[0089] max|Slope(k-1,k)-Slope(k,k+1)|.

[0090] Among them, Slope(k-1,k) represents the slope of the line connecting k-1 and k; Slope(k,k+1) represents the slope of the line connecting k and k+1.

[0091] ⑤ When determining the optimal value of k, the final distance between the sub-region vector and the k center vector is calculated using the weight and the optimal value of the k center vectors.

[0092] ⑥ Determine the echo category of each sub-region vector according to the final distance between all sub-region vectors and the k center vectors when k takes the optimal value, and determine the sub-region vectors with the same type of echo as a cluster in the clustering result.

[0093] After determining the value of k, all superpixels are divided into k clusters according to the results of k-means clustering. At this time, there may be two cases for the value of k: k = 3 and k ≥ 4. The clusters need to be further optimized according to the image.

[0094] Step 106: Determine the echo detection result of each sub-region according to the area of ​​each cluster in the clustering result.

[0095] Step 106 implements cluster optimization, which specifically includes:

[0096] If k=3, the three clusters are optimized to obtain the optimized clusters, and the optimized clusters are used as the echo detection results of each sub-region. If k≥4, the cluster with the smallest area is merged into the cluster closest to the cluster until k=3, and the three clusters are optimized again to obtain the optimized clusters, and the optimized clusters are used as the echo detection results of each sub-region.

[0097] The process of optimizing the three clusters includes: determining the three clusters as a low-echo cluster, an isoechoic cluster and a high-echo cluster in the order of average pixel values ​​from low to high; merging the clusters whose areas among the low-echo cluster, the isoechoic cluster and the high-echo cluster are lower than a set proportion of the area of ​​the region of interest into the cluster closest to the cluster to obtain the optimized cluster.

[0098] For example, the set ratio may be 20%. Figure 5 As shown in the figure, if k>3, the cluster with the smallest area is merged into the cluster with the closest distance until k=3; the three clusters correspond to the low echo cluster, the equal echo cluster, and the high echo cluster according to the average pixel value from low to high; if the area of ​​any cluster among the low echo cluster, the equal echo cluster, and the high echo cluster is less than 20% of the area of ​​the region of interest, the cluster is merged into the cluster with the closest distance to obtain the optimized cluster, that is, the final echo result.

[0099] A specific implementation process of the above-mentioned carotid artery plaque echo detection method is as follows: Figure 6 shown.

[0100] This embodiment uses computers and algorithms to assist doctors in determining the characteristics of plaques in ultrasound images. According to the region of interest of the input carotid artery plaque ultrasound image, the image is processed and segmented based on the superpixel algorithm, the vector space is constructed through mapping, and the carotid artery plaque echo is analyzed and the echo results are output through the unsupervised learning algorithm. The echo results are divided into three categories: low echo, equal echo, and high echo. This embodiment can quickly analyze the echo and include it in the report, which can greatly improve the efficiency of diagnosis and treatment and reduce the pressure on doctors.

[0101] This embodiment has the following advantages:

[0102] (1) Based on unsupervised learning, it is not affected by the global parameters of the image (such as overall brightness, contrast, and resolution) and has strong generalization ability.

[0103] (2) The vector space is constructed using distribution, skewness, and kurtosis instead of the traditional brightness, contrast, etc., which greatly reduces the impact of sample discrete values, outliers, and extreme values.

[0104] (3) The improved k-means algorithm has higher accuracy and stronger robustness than the traditional k-means algorithm.

[0105] (4) No training data is required, and all parameters are adaptively optimized based on the input image.

[0106] Example 2

[0107] A computer device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the carotid plaque echo detection method in Example 1.

[0108] Example 3

[0109] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the carotid plaque echo detection method in Example 1.

[0110] Example 4

[0111] A computer program product includes a computer program, which implements the steps of the carotid plaque echo detection method in Example 1 when executed by a processor.

[0112] Example 5

[0113] A computer device whose internal structure diagram can be shown as Figure 7As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store pending transactions. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the carotid plaque echo detection method in Example 1 is implemented.

[0114] It should be noted that the object information (including but not limited to object device information, object personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the present invention are all information and data authorized by the object or fully authorized by all parties.

[0115] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided by the present invention may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited thereto. The processor involved in each embodiment provided by the present invention may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited thereto.

[0116] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0117] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for detecting carotid artery plaque echo, characterized in that: The method comprises: Obtain ultrasound images of carotid artery plaques; Segmenting the region of interest of the carotid plaque ultrasound image into a plurality of sub-regions; Calculating statistics of each of the sub-regions; the statistics include: distribution, skewness and kurtosis; constructing a sub-region vector for each of the sub-regions according to the statistics; The sub-region vectors of all sub-regions are clustered using an improved k-means algorithm to obtain a clustering result; the improved k-means algorithm is a k-means algorithm that adds weights when calculating the distance between the sub-region vector and the center vector; the center vector is randomly generated; the clustering result includes k clusters, each of which includes sub-region vectors with similar echoes; Determining the echo detection result of each sub-area according to the area of ​​each cluster in the clustering result; Determining the echo detection result of each sub-region according to the area of ​​each cluster in the clustering result specifically includes: If k=3, the three clusters are optimized to obtain optimized clusters, and the optimized clusters are used as the echo detection results of each of the sub-regions; If k≥4, the cluster with the smallest area is merged into the cluster closest to the cluster until k=3, and then the three clusters are optimized to obtain an optimized cluster, and the optimized cluster is used as the echo detection result of each sub-region; The process of optimizing the three clusters includes: The three clusters are identified as a low-echo cluster, an isoechoic cluster, and a high-echo cluster in the order of average pixel values ​​from low to high; The clusters whose areas are lower than a set proportion of the area of ​​the region of interest among the low-echo cluster, the isoechoic cluster and the high-echo cluster are merged into the cluster closest to the cluster to obtain an optimized cluster.

2. The method for detecting carotid artery plaque echo according to claim 1, characterized in that: The improved k-means algorithm is used to cluster the sub-region vectors of all sub-regions to obtain clustering results, including: Randomly generate k center vectors with different k values; where k = 3, 4, 5, ...; For any value of k, the distance between the matrix elements of each sub-region vector and the corresponding matrix elements of the k center vectors is calculated, and a weight is assigned to the distance between each matrix element of the sub-region vector and the corresponding matrix element of the k center vectors to obtain the final distance between the sub-region vector and the k center vectors; each parameter in the statistic is taken as a matrix element; For any k value, according to the final distance, with the goal of minimizing the partial square sum of errors, the values ​​of the center vector and the weight are optimized cyclically and alternately to obtain the optimal values ​​of the weight and k center vectors; The value of k is changed, and the optimal value of k is determined by the elbow point method; the optimal value is the k value corresponding to the inflection point in the k-SSE line graph; the k-SSE line graph represents the relationship between different k values ​​and the minimum partial sum of squares of error; When determining the optimal value of k, the final distance between the sub-region vector and the k center vectors is calculated using the weight and the optimal values ​​of the k center vectors; The echo category of each sub-region vector is determined according to the final distance between all sub-region vectors and the k center vectors when k takes the optimal value, and the sub-region vectors with the same type of echo are determined as a cluster in the clustering result.

3. The method for detecting carotid artery plaque echo according to claim 2, characterized in that: If the matrix element is a distribution, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Wasserstein distance; if the matrix element is a skewness or a kurtosis, the distance between the matrix element of the sub-region vector and the corresponding matrix elements of the k center vectors is the Euclidean distance.

4. The method for detecting carotid artery plaque echo according to claim 1, characterized in that: The region of interest of the carotid artery plaque ultrasound image is divided into a plurality of sub-regions, specifically including: The region of interest of the carotid artery plaque ultrasound image is segmented into multiple sub-regions using a superpixel algorithm.

5. The method for detecting carotid artery plaque echo according to claim 1, characterized in that: The set ratio is 20%.

6. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the carotid plaque echo detection method according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the carotid plaque echo detection method described in any one of claims 1 to 5 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the carotid plaque echo detection method described in any one of claims 1 to 5 are implemented.

Images