A method and system for differentiating between mastitis with abscess

Abstract

The application relates to a method and system for identifying inflammatory mass stage plasma cell mastitis, wherein the method comprises the following steps: S1, acquiring a gray-scale ultrasound image and an ultrasound contrast image of a breast; S2, extracting radiomics features of the gray-scale ultrasound image and the ultrasound contrast image of the breast; and S3, predicting the radiomics features by using a preset model to determine whether the breast has inflammatory mass stage plasma cell mastitis. The application can effectively identify inflammatory mass stage plasma cell mastitis and has good practicability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of breast diagnosis, in particular to a method and system for identifying inflammatory mass phase plasma mastitis. BACKGROUND

[0002] The prior art can construct an inflammatory mass phase plasma mastitis differential diagnosis model based on imaging features. Non-invasively obtain image features that cannot be identified by the naked eye from medical images such as CT, MRI, and ultrasound images through automatic or semi-automatic software and convert them into quantitative features. Through deep data mining, a model with diagnostic, prognostic or predictive value is constructed using machine learning algorithms, which provides information for personalized diagnosis, sub-layering, prediction and disease management. The main process is as follows: first, extract medical features: medical image acquisition; draw / segment the part of interest (lesion) in the image (ROI); use automatic or semi-automatic software to extract the imaging features of the segmented part. Second, use various methods to reduce the dimensionality and feature selection of the extracted imaging features. Finally, use a single machine learning algorithm such as logistic regression to construct a model; the model results can be further combined with clinical information to construct a joint model for the diagnosis of plasma mastitis.

[0003] In the aspect of imaging feature extraction, previous studies mostly use CT and MRI images, and the data is relatively single. The prior art mostly uses single machine learning algorithms such as logistic regression and support vector machine to construct diagnostic models, which may have problems such as overfitting and poor generalization ability, and cannot effectively diagnose mastitis. SUMMARY

[0004] Therefore, the technical problem to be solved by the present application is to overcome the problem that the method for detecting inflammatory mass phase plasma mastitis in the prior art cannot effectively diagnose mastitis due to the use of a single machine learning algorithm, which may cause problems such as overfitting and poor generalization ability.

[0005] To solve the above technical problems, the present application provides a method for identifying inflammatory mass phase plasma mastitis, comprising:

[0006] Step S1: obtaining a gray-scale ultrasound image and an ultrasound contrast image of the breast;

[0007] Step S2: extracting imaging features of the gray-scale ultrasound image and the ultrasound contrast image of the breast;

[0008] Step S3: predicting the imaging features by a preset model to determine whether the breast has inflammatory mass phase plasma mastitis.

[0009] In one embodiment of the present invention, the preset model in step S3 includes a first-level learner and a second-level learner, the first-level learner includes at least two basic learners, and the second-level learner is a meta-learner, wherein,

[0010] The prediction results obtained by independently training each of the basic learners are combined into a new data set, and the new data set is used as the input of the meta-learner to train the meta-learner to obtain a preset model.

[0011] In one embodiment of the present invention, when the new dataset is used as input of the meta-learner, the method further includes:

[0012] First, the K-fold cross-validation method is used to divide the imaging omics features into K data subsets. For each data subset, the K-fold cross-validation method is used to divide the imaging omics features into K data subsets. The data is used as the training set, and the rest The data is used as the validation set;

[0013] Secondly, for the training set in each data subset, after the basic learner is trained with the training set each time, the basic learner is used to predict the corresponding validation set in the data subset to obtain the corresponding prediction results. All the prediction results corresponding to the K data subsets are used as a new data set to train the meta-learner to obtain the preset model.

[0014] In one embodiment of the present invention, different weights are set for the prediction results obtained for each data subset, and the prediction results are multiplied by the corresponding weights to obtain a new data set for input into the meta-learner. The formula is:

[0015]

[0016] in, The prediction results calculated by cross-validation for each base learner; is the weight of each basic learner, all weights are non-negative and the sum is 1; is the new dataset used to input the meta-learner.

[0017] The weights of each basic learner Make adjustments, including:

[0018] In one embodiment of the present invention, when the weight of the base learner When the value is between 0 and 0.3, the first weight formula is used to increase the weight of the basic learner. Size, where the first weight formula is:

[0019] ,in, is the assignment symbol;

[0020] when the weight of the base learner is greater than 1.0 between 0.7 and 1.0, the weight of the base learner is reduced by a second weight formula , wherein the second weight formula is:

[0021] , wherein, is a constant coefficient and ;

[0022] when the weight of the base learner is greater than 1.0 between 0.4 and 0.6, the weight of the base learner is unchanged;

[0023] if the sum of the adjusted weights of the base learners is not 1, the weight of the base learner between 0.4 and 0.6 is increased or decreased to ensure that the sum of the weights of the base learners is 1.

[0024] In an embodiment of the present application, the imageomics features are input into each base learner to obtain corresponding prediction results, and the final prediction result of the meta-learner is obtained by combining the weights calculated in the formula , and the formula is:

[0025]

[0026] , wherein, is a prediction value obtained by using the imageomics features through each base learner; is the final prediction value of the meta-learner.

[0027] In an embodiment of the present application, the step S3 further comprises: performing feature dimension reduction on the imageomics features by LASSO feature screening or principal component analysis.

[0028] In an embodiment of the present application, the meta-learner is any one of a support vector machine, a logistic regression, a random forest, and an artificial neural network; and the base learner is any one of a K-nearest neighbor algorithm, a support vector machine, a logistic regression, and a naive Bayes.

[0029] To solve the above technical problems, the present application provides a granulomatous mastitis identification system, comprising:

[0030] an acquisition module configured to acquire a gray-scale ultrasound image and an ultrasound contrast image of a breast;

[0031] an extraction module configured to extract imageomics features of the gray-scale ultrasound image and the ultrasound contrast image of the breast; ​​​​

[0032] The prediction module is configured to predict the radiomics features by using a preset model to determine whether the breast has inflammatory mass phase plasma breast inflammation.

[0033] To solve the above technical problems, the present application provides an electronic 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 inflammatory mass phase plasma breast inflammation identification method.

[0034] To solve the above technical problems, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the above-mentioned inflammatory mass phase plasma breast inflammation identification method.

[0035] The above technical solution of the present application has the following advantages compared with the prior art:

[0036] The inflammatory mass phase plasma breast inflammation identification method of the present application constructs an efficient model, which can learn how to best combine the prediction results of different models to improve the prediction performance of the model through the hierarchical process combined with the prediction results of each basic learner;

[0037] The model constructed by the present application for identifying inflammatory mass phase plasma breast inflammation has strong generalization ability and high robustness, and has been proved to have good detection effect through experiments, and can be widely applied in practice. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to make the content of the present application more easily understood, the present application will be further described in detail below according to specific embodiments of the present application and in conjunction with the accompanying drawings.

[0039] Figure 1 is a flowchart of the method of the present application;

[0040] Figure 2 is a schematic diagram of the construction of a preset model in the embodiment of the present application, taking 5-fold cross-validation as an example;

[0041] Figure 3 is a schematic diagram of the gray-scale ultrasound and ultrasound contrast images of the inflammatory mass phase plasma breast inflammation patients in the embodiment of the present application;

[0042] Figure 4 is a calibration chart of the final model in the training set and the validation set in the embodiment of the present application;

[0043] Figure 5 is a decision curve chart of the final model in the training set and the validation set in the embodiment of the present application;

[0044] Figure 6It is the nomogram of the final model in the embodiment of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described below with reference to the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments are not intended to limit the present invention.

[0046] Example 1

[0047] Reference Figure 1 As shown, the present invention relates to a method for identifying plasma cell mastitis in the inflammatory stage, comprising:

[0048] Step S1: Acquire a grayscale ultrasound image and an ultrasound contrast-enhanced image of the breast;

[0049] Step S2: extracting radiomics features of the grayscale ultrasound image and the ultrasound contrast-enhanced image of the breast;

[0050] Step S3: predicting the imaging genomics features using a preset model to determine whether the breast has plasma cell mastitis in the inflammatory stage.

[0051] The following is a detailed description of this embodiment

[0052] First, ultrasound imaging features are extracted from grayscale ultrasound images and ultrasound contrast images. After dimensionality reduction (for example, LASSO feature screening or principal component analysis can be used) and screening, the features are used for model construction.

[0053] The reason why grayscale ultrasound images and ultrasound contrast imaging images are used in this embodiment is that: when examining breast-related diseases, ultrasound images are the main examination method, which has the characteristics of being radiation-free, real-time, low-cost, non-invasive, and highly repeatable. It can obtain corresponding high-quality data without increasing the burden on patients.

[0054] Secondly, this embodiment constructs a preset model comprising a two-layer structure: (1) two or more basic learners (first-level learners); and (2) a meta-learner (second-level learner). The prediction results obtained by the independent training data of each basic learner in the first layer constitute a new dataset; this new dataset serves as the input of the second layer and is used to train the meta-learner to obtain the preset model.

[0055] Since directly using the prediction results generated by the base learner as the input of the meta-learner may lead to model overfitting, this embodiment uses the K-fold cross-validation method to first divide the radiomics features into K data subsets on average. For each data subset, and each time using The data is used as the training set, and the rest data as a validation set. Secondly, for each training set in the data subset, when training a base learner, the validation set is predicted by the base learner to obtain the corresponding prediction result, and after repeating K times in turn, all the prediction results are used as a new data set to train the meta-learner to obtain the preset model. This embodiment takes 5-fold cross-validation as an example, and the specific process is shown in Figure 2

[0056] The embodiment can select multiple algorithms with small correlation between learners and little difference in performance as base learners. The meta-learner can select any algorithm, such as support vector machine (SVM), logistic regression (LR), random forest (RF), artificial neural unit, etc. Among them, the method of using K-fold cross-validation to construct the best weighted combination of prediction results from each base learner is also called “super learner”. Taking “super learner” as an example, the main steps of the preset model are as follows:

[0057] (1) The prediction results obtained by each base learner through cross-validation are combined together to form a new data set.

[0058] (2) The new data set is input into the meta-learner, and the meta-learner calculates the weight of each base learner according to a “standard” to achieve the best performance of the model. The “standard” is used to define and quantify the performance of the algorithm in predicting or explaining data methods, which can be defined by the user. For example: in regression analysis, the “standard” can be set as the minimum mean square error between the prediction result of the model and the actual outcome; in binary classification analysis, the “standard” can be set as the area under the receiver operating characteristic (ROC) curve, that is, the value of 1-AUC (Area under curve) is the minimum. ​

[0059] Different weights are set for the prediction results obtained from each data subset, and the prediction results are multiplied by the corresponding weights to obtain a new data set for input into the meta-learner. The formula is:

[0060] (1)

[0061] in, The prediction results calculated by cross-validation for each base learner; The weights of each base learner when the "standard" is met are all non-negative and the sum is 1; It is the weighted prediction value of each base learner calculated based on the prediction result dataset, that is, the new dataset used to input the meta learner.

[0062] It is worth mentioning that the weights of each basic learner in this embodiment are Make adjustments, including:

[0063] (1) When the weight of the basic learner When the value is between 0 and 0.3, the first weight formula is used to increase the weight of the basic learner. Size, where the first weight formula is:

[0064]

[0065] in, is the assignment symbol;

[0066] (2) When the weight of the basic learner When the value is between 0.7 and 1, the second weight formula is used to reduce the weight of the basic learner. Size, where the second weight formula is:

[0067]

[0068] in, is a constant coefficient and ;

[0069] (3) When the weight of the basic learner When the weight of the base learner is between 0.4 and 0.6, constant.

[0070] The reason why the higher weights are adjusted lower and the lower weights are adjusted higher in this embodiment is that it can more effectively prevent the model from overfitting in the future, making the model of this embodiment more generalizable.

[0071] It should be noted that if the weight of the adjusted basic learner If the sum of the weights of the base learners is not equal to 1, the weights of the base learners between 0.4 and 0.6 are increased or decreased to ensure that the weights of the base learners are equal to 1 , specifically:

[0072] (a) If the sum of the weights of the base learners is less than 0.4 or greater than 0.6, the weights of the base learners between 0.4 and 0.6 are increased or decreased by the number of base learners between 0.4 and 0.6 to make the sum of the weights of the base learners equal to 1 . (b) If the sum of the weights of the base learners is not less than 0.4 or not greater than 0.6, the weights of the base learners are not adjusted, because the adjusted values are too high or too low, which is not conducive to the generalization of the model in the later stage. , the sum of the weights of the base learners after adjustment is equal to 1.

[0073] (b) If the sum of the weights of the base learners is not less than 0.4 or not greater than 0.6, the weights of the base learners are not adjusted, because the adjusted values are too high or too low, which is not conducive to the generalization of the model in the later stage.

[0074] (3) The original data (i.e. radiomics features) are used to obtain the corresponding prediction results of each base learner, and the weights calculated in formula (1) are combined to obtain the final prediction result of the meta-learner:

[0075] (2)

[0076] wherein, is the prediction value obtained by using the original data set (i.e. radiomics features) through each base learner; is the weight of each base learner calculated in formula (1); is the final prediction value calculated by combining the prediction results of each base learner using the original data set (i.e. the final prediction result of the meta-learner).

[0077] ​​​​​​​​​​​​​​In summary, the external test method of the preset model is: (1) using the same basic learner to construct a model based on the entire test set data to obtain a corresponding prediction result; and (2) combining the prediction result with the calculated weight to obtain a final prediction value.

[0078] The Stacking method in the application is a super learner (Super Learner) with K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Logistic Regression (LR) and Naive Bayes (NB) as basic learners. Each basic learner is parameter-optimized, and the specific optimization method is as follows: (1) KNN: find the K value with the highest accuracy through 10-fold cross-validation; (2) SVM: through the method of 10-fold cross-validation, grid search the optimal parameters gamma and cost value; (3) LR: through the stepwise regression algorithm, screen the optimal variables to construct a model; (4) NB: set Laplace to 1. The number of cross-validation (K) is set to 2 times according to the size of the sample amount.

[0079] The application is further described below through a specific case:

[0080] The application provides a method for constructing a differential diagnosis model of inflammatory mass phase plasma cell mastitis, which is used for distinguishing inflammatory mass phase plasma cell mastitis patients from breast invasive ductal carcinoma patients, and comprises the following steps:

[0081] (1) extracting ultrasound imageomics features based on ultrasound images:

[0082] Firstly, two ultrasound images and clinical information of 341 patients with plasma cell mastitis or breast invasive ductal carcinoma are collected when the patients receive gray-scale ultrasound and ultrasound contrast examination. When the patients receive the examination, Mindray Resona7, LOGIQ E9 and MyLab TM ClassC ultrasound instruments are used. According to the different ultrasound instruments used by the patients when receiving the examination, a training set (Mindray Resona7, LOGIQ E9) and a verification set (MyLab TM ClassC) are divided. The training set is used for feature dimension reduction, screening and model construction, and the verification set is used for model performance verification.

[0083] Secondly, an ultrasound physician with many years of ultrasound experience uses ITK-SNAP 3.8.0 software to outline and segment (ROI) the region of interest in the image (for details, see Figure 3 , Figure 3 A and C in the above table respectively represent the gray-scale ultrasound and ultrasound contrast examination images, Figure 3B, D in the formula respectively represent segmentation of a region of interest (ROI) for a lesion), and ultrasound image features in the two images are respectively extracted, dimensionally reduced and screened. The dimension reduction and screening process includes: removing feature columns containing "0"; a selectKbest method selects features with a P value less than 0.05; and finally, a Lasso algorithm is used for further screening to obtain a final feature dataset for model construction. In the present application, 788 features are extracted from the two ultrasound images. After dimension reduction and screening, 14 features are finally obtained from the gray-scale ultrasound image; 11 features are obtained from the ultrasound contrast, and 32 combined features of the two images are obtained, a total of three feature datasets.

[0084] (2) Stacking method for constructing a model:

[0085] Using three feature datasets to build models, two, three or four of the four algorithms of KNN, SVM, LR and NB were used as base learners, and the final result was obtained by stacking the prediction results of the base learners. A total of 33 models were constructed, and the results showed that based on the joint dataset, the model constructed by KNN and SVM as base learners had the best performance. The AUC (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and F1 value of the model in the training set and validation set were: 0.974 (0.956, 0.993), 0.932, 0.942, 0.922, 0.924, 0.941, 0.933; 0.863 (0.802, 0.924), 0.778, 0.714, 0.846, 0.833, 0.733 and 0.769. The model was superior to the single model constructed by using KNN algorithm or SVM algorithm alone. The AUC (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and F1 value of the model constructed by using KNN algorithm alone in the training set and validation set were: 0.882 (0.837, 0.927), 0.816, 0.893, 0.738, 0.773, 0.874, 0.829; 0.882 (0.828, 0.937), 0.741, 0.600, 0.892, 0.857, 0.674 and 0.706. The AUC (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and F1 value of the model constructed by using SVM algorithm alone in the training set and validation set were: 0.985 (0.970, 1.000), 0.971, 0.971, 0.971, 0.971, 0.971, 0.971; 0.806 (0.732, 0.880), 0.733, 0.714, 0.754, 0.758, 0.710 and 0.735. The results showed that the model constructed by stacking KNN and SVM algorithms using Stacking method improved the accuracy of the single model; and in the validation set, the sensitivity and specificity values of the model were relatively balanced. The model was called the best model (SL model).

[0086] (3) Construction of the final model:

[0087] The prediction results of the best model were combined with clinical data to construct the final model. The final model was evaluated from three aspects of discrimination, calibration and clinical utility. Single factor and multi-factor logistic was used to screen variables to construct the final model. The final model included: patient age, maximum lesion diameter, whether the neutrophil count was greater than 6.3 × 10 9The classification results of / L and the best model. The AUC (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and F1 value of the final model on the training set and the validation set were 0.978 (0.961, 0.996), 0.942, 0.961, 0.922, 0.925, 0.960, 0.942; 0.938 (0.902, 0.974), 0.830, 0.814, 0.846, 0.851, 0.809, and 0.832, respectively. The model had good calibration ability and clinical utility. The calibration plot of the model is shown in FIG. 6. Figure 4 Figure 5 The decision curve plot of the model in the training set and the validation set is shown in FIG. 7. The results show that the model has high clinical benefit in both the training set and the validation set. The nomogram of the final constructed model is shown in FIG. 8. Figure 6

[0088] Example Two

[0089] The embodiment provides a granulomatous mastitis identification system in the inflammatory mass phase, comprising:

[0090] An acquisition module is configured to acquire a gray-scale ultrasound image and an ultrasound contrast image of a breast.

[0091] An extraction module is configured to extract radiomics features of the gray-scale ultrasound image and the ultrasound contrast image of the breast.

[0092] A prediction module is configured to predict the radiomics features by using a preset model, so as to determine whether the breast has granulomatous mastitis in the inflammatory mass phase.

[0093] Example Three

[0094] The embodiment provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the granulomatous mastitis identification method in the inflammatory mass phase in the embodiment are implemented.

[0095] Example Four

[0096] The embodiment provides a computer-readable storage medium, which stores a computer program. When the processor executes the computer program, the steps of the granulomatous mastitis identification method in the inflammatory mass phase in the embodiment are implemented.

[0097] ​​Those skilled in the art will appreciate that embodiments of the application can be readily used as software, hardware, or a combination of software and hardware. In one embodiment, the application can be implemented in software and / or firmware. In addition, the software implementation can be implemented by one or more computer programs.

[0098] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart

[0099] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart

[0100] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart

[0101] While the preferred embodiments of the application have been described, additional variations and modifications can be employed. Therefore, the terms and expressions

[0102] Obviously, the above-mentioned embodiments are only examples for clearly illustrating the present application, and are not intended to limit the present application. Based on the above description, other different forms of changes or variations can be made by those skilled in the art. Here, it is not necessary and also impossible to enumerate all the embodiments. The obvious changes or variations derived from the above description are still within the protection scope of the present application.

Claims

1. A method for identifying a granulocytic mastitis in an inflammatory stage, characterized by: The method comprises the following steps: Step S1: acquiring a gray-scale ultrasound image and an ultrasound contrast image of a breast; Step S2: extracting radiomics features of the gray-scale ultrasound image and the ultrasound contrast image of the breast; Step S3: predicting the radiomics features by using a preset model to determine whether the breast has inflammatory mass phase plasma cell mastitis. The preset model in the step S3 comprises a primary learner and a secondary learner, the primary learner comprises at least two basic learners, and the secondary learner is a meta-learner, wherein, The prediction results obtained by independently training each basic learner are combined to form a new data set, and the new data set is used as an input of the meta-learner to train the meta-learner to obtain the preset model; When the new data set is used as the input of the meta-learner, the method further comprises the following steps: First, the radiomics features are divided into K subsets using K-fold cross-validation method. For each subset, the data in the subset is used as the training set, and the rest of the data is used as the validation set. the rest of the data is used as the validation set.​ Secondly, for each training set in each data subset, after training the basic learner by using the training set each time, the corresponding prediction result is obtained by predicting the corresponding validation set in the data subset by using the basic learner, and the prediction results corresponding to the K data subsets are used as a new data set to train the meta-learner to obtain the preset model. Different weights are set for the prediction results obtained for each data subset, and the prediction results are multiplied by the corresponding weights to obtain a new data set used as an input of the meta-learner, and the formula is as follows: ； wherein, are prediction results computed by the individual base learners through cross-validation; are weights of the individual base learners, the weights being non-negative and summing to 1 ; are new data sets used to input the meta-learner; weights of the individual base learners adjusting, the method comprising: when the weight of the base learner is between 0 and 0.3, the weight of the base learner is increased by a first weight formula wherein the first weight formula is: size, wherein the first weight formula is: wherein is an assignment symbol; when the weight of the base learner is between 0.7 and 1, the weight of the base learner is reduced by a second weight formula where the second weight formula is: size, where the second weight formula is: wherein is a constant coefficient and ; When the weight of the base learner is between 0.4-0.6 the weight of the base learner is between 0.4-0.6 unchanged; If the sum of the weights of the adjusted base learners is not 1, then the weights of the base learners are increased or decreased by a value between 0.4 and 0.6 to ensure that the sum of the weights of the base learners is 1.

1. A method for training a machine learning model, the method comprising:​​ 2. The method according to claim 1, wherein the method is for identifying a mastitis in an inflammatory stage. The imaging features are input into each base learner to obtain corresponding prediction results, and the final prediction result of the meta-learner is obtained by combining the weights calculated in the formula ​ ； wherein, is a prediction value obtained by each base learner using the radiomic features; is the final prediction value of the meta-learner.

3. The method according to claim 1, wherein the method is for identifying a mastitis in an inflammatory stage. The meta-learner is any one of a support vector machine, a logistic regression, a random forest and an artificial neural network, and the basic learner is any one of a K nearest neighbor algorithm, a support vector machine, a logistic regression and a naive Bayes.

4. A granulomatous mastitis discriminating system for implementing the granulomatous mastitis discriminating method according to any one of claims 1 to 3, characterized by: The method comprises the following steps: An acquisition module is configured to acquire a gray-scale ultrasound image and an ultrasound contrast image of a breast; An extraction module is configured to extract radiomics features of the gray-scale ultrasound image and the ultrasound contrast image of the breast; A prediction module is configured to predict the radiomics features by using a preset model to determine whether the breast has inflammatory mass phase plasma cell mastitis.

5. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the steps of the method for identifying inflammatory mass phase plasma cell mastitis according to any one of claims 1 to 3.

6. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to implement the steps of the method for identifying inflammatory mass phase plasma cell mastitis according to any one of claims 1 to 3.

Images