Training method, grading method and device for arterial plaque neovascularization grading model

Abstract

The present application discloses a training method, a grading method and a device for an arterial plaque neovascularization grading model, wherein the method comprises: after obtaining a sample arterial plaque image and a corresponding arterial plaque neovascularization grading result, constructing a first feature set by image features of the sample arterial plaque image, performing a significance check on the second feature set to obtain a second feature set, and performing feature selection on the second feature set to obtain a third feature set, and training an initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result, thereby improving the accuracy of the grading result of the trained target arterial plaque neovascularization grading model.

Description

Technical Field

[0001] The present application relates to the field of medical imaging, and specifically to a training method, a grading method and a device for an arterial plaque neovascularization grading model. Background Art

[0002] Atherosclerosis (AS) is a systemic, chronic inflammatory disease characterized by lipid accumulation in the arterial intima and local fibrosis, which leads to the formation of stable and unstable plaques. During the progression of atherosclerosis, the thickness of the arterial intima-media thickens, and new blood vessels are formed at the bottom and edge of the plaque. These new blood vessels are prone to rupture due to the lack of a complete basement membrane, which can cause internal bleeding in the plaque or rupture of the fibrous cap, significantly increasing the risk of cardiovascular and cerebrovascular events.

[0003] In the related art, the arterial plaque image is usually manually analyzed to obtain the neovascularization grading result of the arterial plaque image. However, the efficiency of obtaining the neovascularization grading result is relatively low. Therefore, how to obtain an arterial plaque neovascularization grading model is a technical problem that is urgently needed for conveniently grading the neovascularization results of arterial plaque images. Summary of the invention

[0004] To this end, one purpose of the present application is to propose a training method, a grading method and a device for an arterial plaque neovascularization grading model.

[0005] According to an embodiment of one aspect of the present application, a training method for an arterial plaque neovascularization grading model is proposed, the method comprising: obtaining a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image; performing image feature extraction on the sample arterial plaque image to obtain multiple image features of the sample arterial plaque image; constructing a first feature set based on the multiple image features, and performing a significance test on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant effect on the arterial plaque neovascularization grading result; training an initial arterial plaque neovascularization grading model based on the second feature set and the arterial plaque neovascularization grading result to obtain an importance score of each feature in the second feature set; performing feature selection on the second feature set based on the importance score to obtain a third feature set; and retraining the initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model.

[0006] According to another aspect of the present application, an embodiment of an arterial plaque neovascularization grading method is proposed, comprising: obtaining a target arterial plaque image of an object to be processed; constructing a first target feature set based on multiple image features extracted from the arterial plaque image; determining, from the first target feature set, a second target feature set to be input into the target arterial plaque neovascularization grading model, wherein the target arterial plaque neovascularization grading model is obtained by the arterial plaque neovascularization grading method disclosed in an embodiment of the present application; inputting the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

[0007] According to another aspect of the present application, an embodiment provides a training device for an arterial plaque neovascularization grading model, the device comprising: a first acquisition module, used to acquire a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image; a feature extraction module, used to perform image feature extraction on the sample arterial plaque image to obtain a plurality of image features of the sample arterial plaque image; a significance check module, used to construct a first feature set based on the plurality of image features, and perform a significance check on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant effect on the arterial plaque neovascularization grading result; a first training module, used to train an initial arterial plaque neovascularization grading model based on the second feature set and the arterial plaque neovascularization grading result to obtain an importance score of each feature in the second feature set; a feature selection module, used to perform feature selection on the second feature set based on the importance score to obtain a third feature set; and a second training module, used to retrain the initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model.

[0008] According to another aspect of the present application, an embodiment of an arterial plaque neovascularization grading device is proposed, comprising: a second acquisition module, used to acquire a target arterial plaque image of an object to be processed; a construction module, used to construct a first target feature set based on multiple image features extracted from the arterial plaque image; a determination module, used to determine, from the first target feature set, a second target feature set to be input into the target arterial plaque neovascularization grading model, wherein the target arterial plaque neovascularization grading model is obtained by the training method of the arterial plaque neovascularization grading model disclosed in the embodiment of the present application; a grading module, used to input the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

[0009] According to another aspect of the present application, an embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the training method for the arterial plaque neovascularization grading model disclosed in the embodiment of the present application, or the arterial plaque neovascularization grading method is implemented.

[0010] According to another aspect of the present application, an embodiment provides a computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, it is used to implement the training method of the arterial plaque neovascularization grading model disclosed in the embodiment of the present application, or the arterial plaque neovascularization grading method.

[0011] According to another aspect of the present application, an embodiment provides a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, it implements the training method of the arterial plaque neovascularization grading model disclosed in the embodiment of the present application, or the arterial plaque neovascularization grading method.

[0012] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0013] After obtaining a sample arterial plaque image and a corresponding arterial plaque neovascularization grading result, a first feature set is constructed based on the image features of the sample arterial plaque image, and a significance check is performed on the second feature set to obtain a second feature set, and feature selection is performed on the second feature set to obtain a third feature set, and based on the third feature set and the arterial plaque neovascularization grading result, an initial arterial plaque neovascularization grading model is trained, which can improve the accuracy of the grading results of the trained target arterial plaque neovascularization grading model.

[0014] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0016] Figure 1 A schematic flow chart of a method for training an arterial plaque neovascularization grading model provided in an embodiment of the present application.

[0017] Figure 2 A detailed flowchart of step 102 provided in an embodiment of the present application.

[0018] Figure 3 A schematic flow chart of a method for grading arterial plaque neovascularization provided in an embodiment of the present application.

[0019] Figure 4 A schematic diagram of the structure of a training device for an arterial plaque neovascularization grading model provided in an embodiment of the present application.

[0020] Figure 5 A schematic diagram of the structure of an arterial plaque neovascularization grading device provided in an embodiment of the present application.

[0021] Figure 6 It is a schematic diagram of the structure of an electronic device shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0022] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0023] The following describes a training method, a grading method, and a device for an arterial plaque neovascularization grading model according to an embodiment of the present application with reference to the accompanying drawings.

[0024] Figure 1 A schematic flow chart of a method for training an arterial plaque neovascularization grading model provided in an embodiment of the present application.

[0025] like Figure 1 As shown, the method comprises the following steps:

[0026] Step 101, obtaining a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image.

[0027] It should be noted that the training method of the arterial plaque neovascularization grading model in the present embodiment is executed by a training device for the arterial plaque neovascularization grading model, and the training device for the arterial plaque neovascularization grading model can be implemented by software and / or hardware, wherein the training device for the arterial plaque neovascularization grading model can be an electronic device, and can also be configured in an electronic device.

[0028] Among them, the electronic device can be any device with computing capabilities, such as a personal computer, a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, and other hardware devices with various operating systems, touch screens and / or display screens.

[0029] In some embodiments, an ultrasound image of a target artery region of a sample object is acquired, and arterial plaque identification is performed on the ultrasound image to obtain an arterial plaque image region, and the arterial plaque image region is used as a sample arterial plaque image, wherein the arterial plaque image region is the image region where the arterial plaque is located in the ultrasound image.

[0030] It should be noted that the arterial plaque neovascularization grading results of the sample arterial plaque images are determined manually.

[0031] The sample object is a virtual object used as a sample, wherein the virtual object is a virtual object created for a medical object in the real world, and the medical object may be a patient (human) or other animals, such as a cat, a dog, etc. The virtual object may be a patient identity identifier, which may be a patient account, a patient ID, a patient name, etc.

[0032] In this example, a patient is taken as an example for illustrative description.

[0033] It should be noted that the ultrasound image in this embodiment may be various types of ultrasound images. For example, the ultrasound image in this embodiment may be a B-type ultrasound image.

[0034] In this embodiment, after acquiring the ultrasound image, arterial plaque recognition may be performed on the ultrasound image to obtain an image region where the arterial plaque is located in the ultrasound image, and the image region may be segmented to obtain a sample arterial plaque image.

[0035] It should be noted that in different application scenarios, the methods of performing arterial plaque recognition on ultrasound images to obtain the image region where the arterial plaque is located in the ultrasound image are different. For example, arterial plaque recognition can be performed on ultrasound images using a preset arterial plaque recognition algorithm to obtain the image region where the arterial plaque is located in the ultrasound image. For another example, arterial plaque recognition can be performed on ultrasound images using a pre-trained arterial plaque recognition model to obtain the image region where the arterial plaque is located in the ultrasound image.

[0036] It should be noted that the above is only an example of performing arterial plaque recognition on an ultrasound image to obtain an image region where the arterial plaque is located in the ultrasound image. Arterial plaque recognition on an ultrasound image can also be performed by other methods in the relevant technology to obtain an image region where the arterial plaque is located in the ultrasound image. This embodiment does not specifically limit the manner in which arterial plaque recognition is performed on an ultrasound image to obtain an image region where the arterial plaque is located in the ultrasound image.

[0037] It should be noted that the arterial plaque neovascularization grading result in this embodiment may include: whether there are neovascularizations. Correspondingly, if there are neovascularizations, it may also include: neovascularization grading.

[0038] Among them, the neovascularization grading can include: grade 1, grade 2, and grade 3, etc.

[0039] Step 102: extract image features from the sample arterial plaque image to obtain a plurality of image features of the sample arterial plaque image.

[0040] It should be noted that, if the above-mentioned multiple image features are classified, in this embodiment, the above-mentioned multiple image features can be divided into three categories of image features, and the three categories of image features can include: fractal features, imaging genomics features, and deep neural network features.

[0041] The fractal features include one or more of fractal dimension, information dimension, correlation dimension, self-similarity ratio, Shannon entropy, standard index, fractal texture scale and roughness.

[0042] Among them, the radiomics features include one or more of gray level co-occurrence matrix (GLCM), shape features, gray level dependency matrix (GLRLM), gray level size zone matrix (GLSZM), gray level difference matrix (GLDM) and neighborhood gray level difference matrix (NGTDM).

[0043] Among them, Gray Level Co-occurrence Matrix (GLCM): used to describe the gray-level spatial relationship.

[0044] The shape feature includes one or more of the area, perimeter, circularity, and aspect ratio of the patch, which are used to describe the shape.

[0045] Among them, Gray Level Dependence Matrix (GLRLM): is used to describe the distribution of the same gray value in the image.

[0046] Among them, the grayscale size zone matrix (GLSZM) is used to describe the size and distribution of areas with the same grayscale level.

[0047] Among them, Gray Level Difference Matrix (GLDM): is used to describe the differences between different gray levels.

[0048] Among them, the neighborhood grayscale difference matrix (NGTDM) is used to describe the subtle features of grayscale changes.

[0049] In some embodiments, the sample arterial plaque image may be input into a pre-trained deep neural network to perform feature extraction on the sample arterial plaque image through the pre-trained deep neural network, and the output of one or more layers of the deep neural network may be used as deep neural network features.

[0050] Step 103, constructing a first feature set based on multiple image features, and performing a significance test on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant impact on the arterial plaque neovascularization grading result.

[0051] In some embodiments, in order to further improve the accuracy of the subsequently trained target arterial plaque neovascularization grading model, the hemodynamic characteristics in the arterial plaque image area can also be obtained, and the health data of the sample object can be obtained. Correspondingly, before performing a significance test on the first feature set to obtain the second feature set, at least one of the hemodynamic characteristics and the health data can also be added to the first feature set.

[0052] As an example, health data may be added to the first feature set.

[0053] As another example, hemodynamic features may be added to the first feature set.

[0054] As another example, both hemodynamic features and health data may be added to the first feature set.

[0055] In some examples, when the sample arterial plaque image is a Doppler ultrasound image, the hemodynamic characteristics of the sample arterial plaque image can be obtained by extracting the hemodynamic characteristics of the sample arterial plaque image.

[0056] Among them, the hemodynamic characteristics may include but are not limited to blood flow velocity, blood flow volume and blood flow shear stress.

[0057] Among them, it can be understood that the above-mentioned hemodynamic characteristics can be used to more accurately evaluate the activity and rupture risk of new blood vessels.

[0058] Among them, the health data in this embodiment may include but is not limited to age, gender, medical history, possible genetic history, etc.

[0059] In the technical solution of the present invention, the collection, storage, use, processing, transmission, provision and disclosure of the health data involved are all carried out with the user's consent, and are in compliance with the relevant laws and regulations and do not violate public order and good morals.

[0060] Step 104: training the initial arterial plaque neovascularization grading model based on the second feature set and the arterial plaque neovascularization grading result to obtain the importance score of each feature in the second feature set.

[0061] Among them, the arterial plaque neovascularization grading model in this embodiment can be a linear regression model, a support vector machine model, a decision tree model, a random forest model, or a neural network model. This embodiment does not specifically limit the arterial plaque neovascularization grading model. For the linear regression model, the absolute value of the regression coefficient is used to measure the importance of the feature. For the support vector machine model, the weight coefficient of the feature on the hyperplane is used to measure the importance of the feature. For the decision tree or random forest, the Gini index of the feature is used to measure the importance of the feature. For the neural network, the SHAP (SHapleyAdditive exPlanations) value is used to measure the importance of the feature.

[0062] The importance score is used to indicate the importance of the corresponding feature in the second feature set to the prediction result of the arterial plaque neovascularization grading model.

[0063] It should be noted that it is understandable that the more important a feature is to the prediction result of the arterial plaque neovascularization grading model, the higher its corresponding importance score.

[0064] In the case that the initial arterial plaque neovascularization grading model is a Least Absolute Shrinkage and Selection Operator (LASSO) regression model, a possible implementation method of training the initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result to obtain the importance score of each feature in the second feature set is: determining the optimal regularization factor of the LASSO regression model; under the condition of the optimal regularization factor, training the LASSO regression model according to the second feature set and the arterial plaque neovascularization grading result to obtain the trained LASSO regression model; obtaining the regression coefficient of each feature in the second feature set from the trained LASSO regression model, wherein the regression coefficient represents the influence of the corresponding feature in the second feature set on the arterial plaque neovascularization grading result; and determining the importance score of each feature in the second feature set according to the regression coefficient.

[0065] As an example, the regression coefficient may be used as the importance score of each feature in the second feature set. Correspondingly, according to the importance score, the second feature set is selected to obtain the third feature set in an implementation manner that: the features in the second feature set whose importance score is not zero may be used to construct the third feature set. In other words, the features in the second feature set whose regression coefficient is not zero may be used to construct the third feature set.

[0066] In this embodiment, the optimal regularization factor λ of the LASSO regression model can be determined by a cross-validation method.

[0067] Step 105 , performing feature selection on the second feature set according to the importance score to obtain a third feature set.

[0068] In this embodiment, in different application scenarios, according to the importance score, the second feature set is selected to obtain the third feature set in different implementation methods, which are exemplarily described as follows:

[0069] As a possible implementation method, the feature with the lowest importance score in the second feature set is deleted, and the step of training the initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result is jumped to obtain the importance score of each feature in the second feature set, until the number of features in the second feature set reaches the preset number of features, and the second feature set currently obtained is used as the third feature set. Thus, through the feature elimination algorithm, feature selection is accurately implemented for the second feature set, and the third feature set is accurately obtained.

[0070] The preset number of features is a number of features that is preset according to actual needs, and this embodiment does not specifically limit this.

[0071] As another possible implementation manner, features whose importance scores are greater than or equal to a preset score threshold may be selected from the second feature set to form a third feature set.

[0072] The preset scoring threshold is a scoring critical value preset according to actual needs, and this embodiment does not make any specific limitation to this.

[0073] Step 106, retraining the initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model.

[0074] In this embodiment, the features in the third feature set are fused to obtain fused features, and the fused features are input into the arterial plaque neovascularization grading model to obtain a neovascularization grading prediction result, and the arterial plaque neovascularization grading model is trained based on the neovascularization grading prediction result and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model.

[0075] In this embodiment, the loss value of the arterial plaque neovascularization grading model can be determined based on the neovascularization grading prediction results and the arterial plaque neovascularization grading results, and the arterial plaque neovascularization grading model is trained based on the loss value until the loss value meets the preset model termination condition to obtain the target arterial plaque neovascularization grading model.

[0076] The preset model end condition can be configured accordingly according to actual needs. For example, the loss value can meet the preset model end condition when the loss value is less than the preset value, or when the change of the loss value tends to be stable, that is, the difference between the loss values ​​corresponding to two or more consecutive trainings is less than the set value, that is, the loss value basically does not change.

[0077] Among them, it can be understood that in the process of training the arterial plaque neovascularization grading model through the training data set, the model parameters of the arterial plaque neovascularization grading model are continuously adjusted according to the loss value of each training. For example, the model parameters of the arterial plaque neovascularization grading model can be adjusted in the direction of decreasing the loss value. When the loss value meets the preset model termination condition, the target arterial plaque neovascularization grading model is obtained.

[0078] In some embodiments, in order to help reduce the complexity of the model and improve the generalization ability of the model, before fusing the features in the third feature set to obtain the fused features, you can also: you can also combine the principal component algorithm to perform dimensionality reduction processing on the third feature set to obtain a new third feature set.

[0079] The following is an exemplary description of the process of performing dimensionality reduction processing on the third feature set in combination with the principal component algorithm to obtain a new third feature set: performing singular value decomposition (SVD) on the third feature set to obtain a singular value sequence and singular vectors corresponding to each singular value in the singular value sequence, wherein the singular value sequence is obtained by sorting multiple singular values ​​obtained by singular value decomposition in order from large to small; when it is determined that the ratio of the sum of the first k singular values ​​in the singular value sequence to the total singular value is greater than a preset ratio threshold, obtaining the singular vectors corresponding to each of the first k singular values ​​to form the first k singular vectors, wherein the total eigenvalue is obtained by summing each singular value in the singular value sequence, wherein k is greater than or equal to 1 and less than N, wherein N represents the total number of singular values ​​in the singular value sequence; projecting the third feature set into the space of the first k singular vectors to obtain a new third feature set. In this way, the feature dimension can be effectively reduced, the computational complexity can be reduced, and the overall variation information of the data can be retained to the maximum extent.

[0080] Among them, the preset ratio threshold is a preset ratio value, for example, the preset ratio threshold may be 90%, 85%, etc., and may be preset according to actual needs. This embodiment does not specifically limit the value of the preset ratio threshold.

[0081] The training method of the arterial plaque neovascularization grading model provided in the embodiment of the present application, after obtaining a sample arterial plaque image and a corresponding arterial plaque neovascularization grading result, constructs a first feature set by image features of the sample arterial plaque image, and performs a significance check on the second feature set to obtain a second feature set, and performs feature selection on the second feature set to obtain a third feature set, and trains the initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result, which can improve the accuracy of the grading results of the trained target arterial plaque neovascularization grading model.

[0082] In some exemplary embodiments, in order to clearly understand the process of performing a significance test on the first feature set to obtain the second feature set, the following is combined with Figure 2 A possible implementation of the above step 102 is described exemplarily.

[0083] Figure 2 A detailed flowchart of step 102 provided in an embodiment of the present application.

[0084] like Figure 2 As shown, it may include:

[0085] Step 201, for any target feature in the first feature set, determine the significance value of the target feature, wherein the significance value is inversely related to the significance of the impact of the target feature on the result of arterial plaque neovascularization grading.

[0086] It should be noted that the inverse relationship in this example means that when the significance of the target feature on the arterial plaque neovascularization grading result is higher, the significance value corresponding to the target feature is smaller, and vice versa.

[0087] Step 202: sort the features in the first feature set in ascending order of significance value to obtain a sorting result.

[0088] Step 203: According to the ranking of the target feature in the sorting result and the total number of features in the first feature set, the preset significance threshold is adjusted to obtain a target significance threshold corresponding to the target feature.

[0089] The total number of features is used to indicate the total number of features in the first feature set.

[0090] It should be noted that the above preset significance threshold is preset according to actual needs. For example, the preset significance threshold α may be 0.05, and this embodiment does not specifically limit the value of the preset significance threshold.

[0091] In addition, the significance threshold preset in this embodiment may also be referred to as a restriction level.

[0092] In this embodiment, the ratio of the ranking of the target feature in the sorting result to the total number of features in the first feature set can be obtained, and the preset significance threshold is weighted according to the ratio to obtain the target significance threshold corresponding to the target feature.

[0093] Among them, the calculation formula for obtaining the target significance threshold α(i) is:

[0094] α(i) = (i / m) × FDR

[0095] Among them, m is the total number of features, i is the rank of the target feature in the sorting results, and FDR represents the preset significance threshold.

[0096] Step 204: construct a second feature set based on the features in the first feature set whose significance values ​​are less than or equal to the target significance threshold.

[0097] It should be noted that when the significance value of the target feature is less than or equal to the target significance threshold, the null hypothesis is determined to be false.

[0098] Among them, the original hypothesis is that the target feature has no effect on the grading results of arterial plaque neovascularization. Among them, the original hypothesis is false, indicating that the target feature has a significant effect on the grading results of arterial plaque neovascularization.

[0099] It can be understood that when the significance value of the target feature is greater than its corresponding target significance threshold, it indicates that the original hypothesis is true. At this time, it can be determined that the target feature has no effect on the arterial plaque neovascularization grading result.

[0100] In this embodiment, features whose significance values ​​are less than or equal to the target significance threshold are obtained from the first feature set, and a second feature set is constructed based on the obtained features. In other words, features that have a significant impact on the arterial plaque neovascularization grading results can be obtained from the first feature set, and a second feature set can be constructed based on the obtained features.

[0101] In this embodiment, by analyzing each target feature in the first feature set, the target significance threshold corresponding to each target feature is determined, and the significance value of the corresponding target feature is compared with its corresponding target significance threshold, and based on the comparison result, it is determined whether the target feature has a significant impact on the arterial plaque neovascularization grading result, and based on the features that have a significant impact on the arterial plaque neovascularization grading result, the second feature set is accurately constructed, thereby reducing the amount of data that needs to be processed during the training process of the arterial plaque neovascularization grading model and the impact of irrelevant features on the arterial plaque neovascularization grading model, and further improving the accuracy of the trained target arterial plaque neovascularization grading model.

[0102] It should be noted that, in addition to using the above-mentioned method to process the features in the second feature set, the following method may also be used to process the features in the second feature set.

[0103] T-test method: For each continuous characteristic variable, such as blood flow velocity, blood flow rate, etc., the t-test can be used to analyze whether the difference in the mean value of the characteristic between different neovascularization grading categories (such as grade 0, grade 1, grade 2) is statistically significant.

[0104] Chi-square test: For categorical characteristic variables, such as gender and whether one has a history of cardiovascular disease, the chi-square test is used to evaluate the correlation between the distribution of characteristic variables and the neovascularization grade.

[0105] Mann-Whitney U test: For continuous features that do not satisfy the normal distribution, the Mann-Whitney U test (non-parametric test) is used to analyze the differences between different classification categories of the feature.

[0106] Correlation analysis: The Pearson correlation coefficient or Spearman rank correlation coefficient is used to evaluate the strength of the correlation between each feature and the neovascularization grading result for simultaneous analysis of multiple features in the first feature set. This analysis can help identify features that are highly correlated with the grading.

[0107] The embodiments of the present application also provide a method for grading arterial plaque neovascularization.

[0108] Figure 3 A schematic flow chart of a method for grading arterial plaque neovascularization provided in an embodiment of the present application.

[0109] Step 301, obtaining a target arterial plaque image of the object to be processed.

[0110] It should be noted that the arterial plaque neovascularization grading method in this embodiment is performed by an arterial plaque neovascularization grading device, which can be implemented by software and / or hardware, wherein the arterial plaque neovascularization grading device can be an electronic device and can also be configured in an electronic device.

[0111] Among them, the electronic device can be any device with computing capabilities, such as a personal computer, a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, and other hardware devices with various operating systems, touch screens and / or display screens.

[0112] The object to be processed refers to a virtual object to be processed.

[0113] For the specific description of the virtual object, please refer to the relevant description in other embodiments, which will not be repeated here.

[0114] In this embodiment, an ultrasound image of the target artery area of ​​the object to be processed can be obtained, and arterial plaque identification can be performed on the ultrasound image to obtain an arterial plaque image area, and the arterial plaque image area is used as the target arterial plaque image, wherein the arterial plaque image area is the image area where the arterial plaque is located in the ultrasound image.

[0115] Step 302: construct a first target feature set based on multiple image features extracted from the arterial plaque image.

[0116] Step 303: Determine a second target feature set from the first target feature set to be input into the target arterial plaque neovascularization grading model.

[0117] It should be noted that the target arterial plaque neovascularization grading model is obtained based on the training method of the arterial plaque neovascularization grading model disclosed in this embodiment.

[0118] Among them, the specific process of training and obtaining the target arterial plaque neovascularization grading model can be found in the relevant descriptions in other embodiments, and will not be repeated here.

[0119] It should be noted that the types of features included in the second target feature set in this embodiment are the same as the types of features included in the third feature set in the training method for the arterial plaque neovascularization grading model disclosed in this embodiment.

[0120] In this embodiment, in order to further improve the accuracy of the subsequent neovascularization grading results determined, the target hemodynamic characteristics in the target arterial plaque image can also be obtained; the target health data of the object to be processed can be obtained; and at least one of the target hemodynamic characteristics and the target health data can be added to the first target feature set.

[0121] It should be noted that the target health data of the object to be processed is similar to the content contained in the health data of the sample object in other embodiments, and this embodiment will not be repeated.

[0122] Step 304: input the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

[0123] It should be noted that corresponding clinical decision support can be provided based on the results of arterial plaque neovascularization grading. For example, for high-risk grade 3 neovascularization, further examination, interventional treatment or drug management can be considered, such as angiography to evaluate the arterial stenosis in detail, and a personalized treatment plan can be formulated based on the patient's overall risk status.

[0124] Based on the above description, it can be seen that this embodiment combines the target arterial plaque neovascularization grading model to process the second target feature set obtained based on the arterial plaque image, and can automatically and accurately determine the arterial plaque neovascularization grading result of the target arterial plaque image, thereby realizing automatic identification and grading of neovascularization, helping to improve the accuracy of atherosclerosis risk assessment and provide effective support for clinical treatment decisions.

[0125] The arterial plaque neovascularization grading method provided in the embodiment of the present application obtains a target arterial plaque image of an object to be processed; constructs a first target feature set based on multiple image features extracted from the arterial plaque image, and determines a second target feature set to be input into a target arterial plaque neovascularization grading model from the first target feature set, and inputs the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image. Thus, the arterial plaque neovascularization grading result of the target arterial plaque image can be accurately determined through the target arterial plaque neovascularization grading model, which is helpful for subsequent diagnosis and treatment based on the results output by the arterial plaque neovascularization grading model.

[0126] Figure 4 A schematic diagram of the structure of a training device for an arterial plaque neovascularization grading model provided in an embodiment of the present application.

[0127] like Figure 4As shown, the training device 40 for the arterial plaque neovascularization grading model may include:

[0128] The first acquisition module 401 is used to acquire a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image.

[0129] The feature extraction module 402 is used to extract image features from the sample arterial plaque image to obtain multiple image features of the sample arterial plaque image.

[0130] The significance check module 403 is used to construct a first feature set based on multiple image features, and perform a significance check on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant impact on the arterial plaque neovascularization grading result.

[0131] The first training module 404 is used to train the initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result to obtain the importance score of each feature in the second feature set.

[0132] The feature selection module 405 is used to perform feature selection on the second feature set according to the importance score to obtain a third feature set.

[0133] The second training module 406 is used to retrain the initial arterial plaque neovascularization grading model according to the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model.

[0134] In one embodiment of the present application, the significance verification module 403 is specifically used to: determine the significance value of the target feature for any target feature in the first feature set, wherein the significance value is inversely related to the significance of the effect of the target feature on the result of arterial plaque neovascularization grading; sort the features in the first feature set in order of significance value from small to large to obtain a sorting result; adjust the preset significance threshold according to the ranking of the target feature in the sorting result and the total number of features in the first feature set to obtain a target significance threshold corresponding to the target feature; construct a second feature set according to the features in the first feature set whose significance values ​​are less than or equal to the target significance threshold.

[0135] In one embodiment of the present application, the feature selection module 405 is specifically used to: delete the feature with the lowest importance score in the second feature set, and jump to the step of training the initial arterial plaque neovascularization grading model based on the second feature set and the arterial plaque neovascularization grading results to obtain the importance score of each feature in the second feature set, until the number of features in the second feature set reaches a preset number of features, and the currently obtained second feature set is used as the third feature set.

[0136] In one embodiment of the present application, when the initial arterial plaque neovascularization grading model is a LASSO regression model, the first training module 404 is specifically used to: determine the optimal regularization factor of the LASSO regression model; under the condition of the optimal regularization factor, train the LASSO regression model according to the second feature set and the arterial plaque neovascularization grading result to obtain a trained LASSO regression model; obtain the regression coefficient of each feature in the second feature set from the trained LASSO regression model, wherein the regression coefficient represents the influence of the corresponding feature in the second feature set on the arterial plaque neovascularization grading result; and determine the importance score of each feature in the second feature set according to the regression coefficient.

[0137] In one embodiment of the present application, the second training module 406 includes:

[0138] A fusion unit, used for fusing the features in the third feature set to obtain a fused feature;

[0139] A prediction unit, used for inputting the fusion features into an arterial plaque neovascularization grading model to obtain a neovascularization grading prediction result;

[0140] The training unit is used to train the arterial plaque neovascularization grading model according to the neovascularization grading prediction result and the arterial plaque neovascularization grading result, so as to obtain a target arterial plaque neovascularization grading model.

[0141] In one embodiment of the present application, before fusing the features in the third feature set to obtain the fused features, the second training module 406 may further include:

[0142] A singular value decomposition unit is used to perform singular value decomposition on the third feature set to obtain a singular value sequence and a singular vector corresponding to each singular value in the singular value sequence, wherein the singular value sequence is obtained by sorting multiple singular values ​​obtained by the singular value decomposition in descending order;

[0143] an acquiring unit, configured to acquire singular vectors corresponding to the first k singular values ​​respectively when it is determined that the ratio of the sum of the first k singular values ​​in the singular value sequence to the total singular value is greater than a preset ratio threshold, wherein the total singular value is obtained by summing up the singular values ​​in the singular value sequence, wherein k is greater than or equal to 1 and less than N, wherein N represents the total number of singular values ​​in the singular value sequence;

[0144] A processing unit, used for forming first k singular vectors according to the singular vectors corresponding to the first k singular values, and projecting the third feature set into the space of the first k singular vectors to obtain a new third feature set;

[0145] The fusion unit is specifically used to fuse the features in the new third feature set to obtain a fused feature.

[0146] In one embodiment of the present application, the device further includes:

[0147] A third acquisition module is used to acquire the hemodynamic characteristics in the sample arterial plaque image;

[0148] A fourth acquisition module is used to acquire health data of the sample object;

[0149] The device may further include:

[0150] An adding module is used to add at least one of the hemodynamic characteristics and the health data to the first feature set.

[0151] It should be noted that the aforementioned explanation of the embodiment of the training method for the arterial plaque neovascularization grading model is also applicable to the training device for the arterial plaque neovascularization grading model of this embodiment, and will not be repeated here.

[0152] The training device for the arterial plaque neovascularization grading model provided in the embodiment of the present application, after obtaining a sample arterial plaque image and a corresponding arterial plaque neovascularization grading result, constructs a first feature set by image features of the sample arterial plaque image, performs a significance check on the second feature set to obtain a second feature set, and performs feature selection on the second feature set to obtain a third feature set, and trains an initial arterial plaque neovascularization grading model based on the third feature set and the arterial plaque neovascularization grading result, thereby being able to improve the accuracy of the grading results of the trained target arterial plaque neovascularization grading model.

[0153] Figure 5 A schematic diagram of the structure of an arterial plaque neovascularization grading device provided in an embodiment of the present application.

[0154] like Figure 5As shown, the arterial plaque neovascularization grading device 500 comprises:

[0155] The second acquisition module 501 is used to acquire a target arterial plaque image of the object to be processed.

[0156] The construction module 502 is used to construct a first target feature set according to a plurality of image features extracted from the arterial plaque image.

[0157] The determination module 503 is used to determine a second target feature set to be input into the target arterial plaque neovascularization grading model from the first target feature set, wherein the target arterial plaque neovascularization grading model is obtained according to the training method of the arterial plaque neovascularization grading model disclosed in the embodiment of the present application.

[0158] The grading module 504 is used to input the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

[0159] Before determining a second target feature set to be input into the target arterial plaque neovascularization grading model from the first target feature set, the apparatus further includes:

[0160] a fifth acquisition module, for acquiring target hemodynamic characteristics in the target arterial plaque image;

[0161] A sixth acquisition module, used to acquire target health data of the object to be processed;

[0162] An adding module is used to add at least one of a target hemodynamic feature and a target health data to the first target feature set.

[0163] It should be noted that the above explanations of the embodiment of the arterial plaque neovascularization grading method are also applicable to the arterial plaque neovascularization grading device of this embodiment, and will not be repeated here.

[0164] The arterial plaque neovascularization grading device provided in the embodiment of the present application obtains a target arterial plaque image of an object to be processed; constructs a first target feature set based on multiple image features extracted from the arterial plaque image, and determines a second target feature set to be input into a target arterial plaque neovascularization grading model from the first target feature set, and inputs the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image. Thus, the arterial plaque neovascularization grading result of the target arterial plaque image can be accurately determined through the target arterial plaque neovascularization grading model, which is helpful for subsequent diagnosis and treatment based on the results output by the arterial plaque neovascularization grading model.

[0165] An embodiment of the present application also provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the training method of the arterial plaque neovascularization grading model of any of the aforementioned embodiments, or the arterial plaque neovascularization grading method.

[0166] Figure 6 It is a schematic diagram of the structure of an electronic device shown in an exemplary embodiment of the present application.

[0167] The present application is described in more detail below using embodiments. The present application can be embodied in a variety of different forms and should not be construed as being limited to the exemplary embodiments described herein.

[0168] like Figure 6 As shown, the electronic device 600 may include: a memory 601, a processor 602, a communication interface 603, and a bus architecture 604, wherein the memory 601 is used to store a computer program for executing the solution of the present application, and the execution is controlled by the processor 602. The processor 602 is used to execute the computer program stored in the memory 601, so as to implement a training method for an arterial plaque neovascularization grading model or an arterial plaque neovascularization grading method provided in the above embodiment of the present application.

[0169] The communication interface 603, the processor 602 and the memory 601 can be connected to each other via a bus architecture 604; the bus architecture 604 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus architecture 604 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0170] The processor 602 may be a CPU, a microprocessor, an ASIC, or one or more integrated circuits for controlling the execution of the program of the present application.

[0171] The communication interface 603 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), wired access networks, etc.

[0172] The memory 601 may be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory may exist independently and be connected to the processor via a bus architecture 604. The memory may also be integrated with the processor.

[0173] The embodiment of the present application also provides a computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, it is used to implement the method disclosed in the embodiment of the present application.

[0174] The embodiment of the present application also provides a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, the method disclosed in the embodiment of the present application is implemented.

[0175] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties. The acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant laws and regulations of the relevant regions.

[0176] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0177] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0178] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0179] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or otherwise processing in a suitable manner if necessary, and then stored in a computer memory.

[0180] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0181] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0182] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0183] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A training method for an arterial plaque neovascularization grading model, characterized in that: The method comprises: Acquire a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image; Performing image feature extraction on the sample arterial plaque image to obtain a plurality of image features of the sample arterial plaque image; Constructing a first feature set according to the multiple image features, and performing a significance test on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant impact on the arterial plaque neovascularization grading result; Training an initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result to obtain an importance score of each feature in the second feature set; performing feature selection on the second feature set according to the importance score to obtain a third feature set; Retraining the initial arterial plaque neovascularization grading model according to the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model; The performing a significance test on the first feature set to obtain a second feature set includes: For any target feature in the first feature set, determining a significance value of the target feature, wherein the significance value is inversely related to the significance of the effect of the target feature on the arterial plaque neovascularization grading result; Sorting the features in the first feature set in ascending order of the significance values ​​to obtain a sorting result; According to the ranking of the target feature in the sorting result and the total number of features in the first feature set, adjusting the preset significance threshold to obtain a target significance threshold corresponding to the target feature; The second feature set is constructed according to features in the first feature set whose significance values ​​are less than or equal to the target significance threshold.

2. The method according to claim 1, characterized in that The performing feature selection on the second feature set according to the importance score to obtain a third feature set includes: The feature with the lowest importance score in the second feature set is deleted, and the process jumps to the step of training the initial arterial plaque neovascularization grading model based on the second feature set and the arterial plaque neovascularization grading result to obtain the importance score of each feature in the second feature set, until the number of features in the second feature set reaches a preset number of features, and the currently obtained second feature set is used as the third feature set.

3. The method according to claim 1, characterized in that In the case where the initial arterial plaque neovascularization grading model is a LASSO regression model, the initial arterial plaque neovascularization grading model is trained according to the second feature set and the arterial plaque neovascularization grading result to obtain an importance score of each feature in the second feature set, including: Determining the optimal regularization factor of the LASSO regression model; Under the condition of an optimal regularization factor, the LASSO regression model is trained according to the second feature set and the arterial plaque neovascularization grading result to obtain a trained LASSO regression model; Obtaining the regression coefficient of each feature in the second feature set from the trained LASSO regression model, wherein the regression coefficient represents the influence of the corresponding feature in the second feature set on the arterial plaque neovascularization grading result; An importance score of each feature in the second feature set is determined according to the regression coefficient.

4. The method according to claim 1, characterized in that: The step of training the arterial plaque neovascularization grading model according to the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model includes: Fusing the features in the third feature set to obtain fused features; Inputting the fusion features into the arterial plaque neovascularization grading model to obtain a neovascularization grading prediction result; The arterial plaque neovascularization grading model is trained according to the neovascularization grading prediction result and the arterial plaque neovascularization grading result to obtain the target arterial plaque neovascularization grading model.

5. The method according to claim 4, characterized in that Before fusing the features in the third feature set to obtain a fused feature, the method further includes: Performing singular value decomposition on the third feature set to obtain a singular value sequence and a singular vector corresponding to each singular value in the singular value sequence, wherein the singular value sequence is obtained by sorting multiple singular values ​​obtained by the singular value decomposition in descending order; When it is determined that the ratio of the sum of the first k singular values ​​in the singular value sequence to the total singular value is greater than a preset ratio threshold, obtaining a singular vector corresponding to each of the first k singular values, wherein the total singular value is obtained by summing up the singular values ​​in the singular value sequence, wherein k is greater than or equal to 1 and less than N, wherein N represents the total number of singular values ​​in the singular value sequence; Forming first k singular vectors according to the singular vectors corresponding to the first k singular values, and projecting the third feature set into the space of the first k singular vectors to obtain a new third feature set; The steps of fusing the features in the third feature set to obtain a fused feature include: The features in the new third feature set are fused to obtain fused features.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Acquiring hemodynamic characteristics in the sample arterial plaque image; Acquiring health data of the sample subject; Before performing a significance test on the first feature set to obtain a second feature set, the method further includes: At least one of the hemodynamic feature and the health data is added to the first feature set.

7. A method for grading arterial plaque neovascularization, characterized in that: include: Acquire a target arterial plaque image of a subject to be treated; constructing a first target feature set according to a plurality of image features extracted from the arterial plaque image; Determining, from the first target feature set, a second target feature set to be input into the target arterial plaque neovascularization grading model, wherein the target arterial plaque neovascularization grading model is obtained according to the method of any one of claims 1-6; The second target feature set is input into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

8. The method according to claim 7, characterized in that The target arterial plaque neovascularization grading model is trained in the following manner: Acquire a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image; Performing image feature extraction on the sample arterial plaque image to obtain a plurality of image features of the sample arterial plaque image; constructing a first feature set according to the plurality of image features; Acquiring hemodynamic characteristics in the sample arterial plaque image; Acquiring health data of the sample subject; adding at least one of the hemodynamic feature and the health data to the first feature set; Performing a significance test on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant impact on the arterial plaque neovascularization grading result; Training an initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result to obtain an importance score of each feature in the second feature set; performing feature selection on the second feature set according to the importance score to obtain a third feature set; Retraining the initial arterial plaque neovascularization grading model according to the third feature set and the arterial plaque neovascularization grading result to obtain a target arterial plaque neovascularization grading model; Before determining a second target feature set to be input into the target arterial plaque neovascularization grading model from the first target feature set, the method further includes: Acquiring target hemodynamic characteristics in the target arterial plaque image; Acquiring target health data of the object to be processed; At least one of the target hemodynamic feature and the target health data is added to the first target feature set.

9. A training device for an arterial plaque neovascularization grading model, characterized in that: The device comprises: A first acquisition module is used to acquire a sample arterial plaque image of a sample object and an arterial plaque neovascularization grading result of the sample arterial plaque image; A feature extraction module, used for performing image feature extraction on the sample arterial plaque image to obtain a plurality of image features of the sample arterial plaque image; a significance check module, configured to construct a first feature set according to the plurality of image features, and perform a significance check on the first feature set to obtain a second feature set, wherein the features in the second feature set have a significant impact on the arterial plaque neovascularization grading result; a first training module, configured to train an initial arterial plaque neovascularization grading model according to the second feature set and the arterial plaque neovascularization grading result, so as to obtain an importance score of each feature in the second feature set; a feature selection module, configured to perform feature selection on the second feature set according to the importance score to obtain a third feature set; a second training module, configured to retrain the initial arterial plaque neovascularization grading model according to the third feature set and the arterial plaque neovascularization grading result, so as to obtain a target arterial plaque neovascularization grading model; The significance verification module is specifically used for: For any target feature in the first feature set, determining a significance value of the target feature, wherein the significance value is inversely related to the significance of the effect of the target feature on the arterial plaque neovascularization grading result; Sorting the features in the first feature set in ascending order of the significance values ​​to obtain a sorting result; According to the ranking of the target feature in the sorting result and the total number of features in the first feature set, adjusting the preset significance threshold to obtain a target significance threshold corresponding to the target feature; The second feature set is constructed according to features in the first feature set whose significance values ​​are less than or equal to the target significance threshold.

10. An arterial plaque neovascularization grading device, characterized in that: include: A second acquisition module is used to acquire a target arterial plaque image of the object to be processed; A construction module, configured to construct a first target feature set according to a plurality of image features extracted from the arterial plaque image; a determination module, configured to determine, from the first target feature set, a second target feature set to be input into the target arterial plaque neovascularization grading model, wherein the target arterial plaque neovascularization grading model is obtained according to the method of any one of claims 1 to 6; A grading module is used to input the second target feature set into the target arterial plaque neovascularization grading model to obtain an arterial plaque neovascularization grading result of the target arterial plaque image.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer program is used to implement the method according to any one of claims 1 to 8.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images