Breast image processing method and device, computer equipment and storage medium

Abstract

The invention belongs to the technical field of artificial intelligence, and relates to a breast image processing method, which is applied to the field of digital medical treatment and comprises the following steps: receiving a to-be-processed breast image; preprocessing the mammary gland image to obtain a corresponding target mammary gland image; extracting a mammary gland area image from the target mammary gland image; calling a pre-trained lump detection model; wherein the lump detection model is a model obtained by training a target area convolutional neural network based on pre-constructed image sample data; performing detection processing on the mammary gland area image based on the lump detection model to obtain a corresponding lump detection result; and outputting the lump detection result. The invention further provides a breast image processing device, computer equipment and a storage medium. In addition, the invention also relates to a block chain technology, and the lump detection result can be stored in a block chain. Based on the use of the lump detection model, the detection precision of lump detection is effectively improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence development technology and the field of digital medicine, and in particular to a method, device, computer equipment and storage medium for processing breast images. Background Art

[0002] Breast cancer is one of the most common and lethal cancers in the world. Early screening and treatment are crucial to improving the survival rate of patients. In the early screening and diagnosis of breast cancer, breast image analysis plays a pivotal role. In particular, rapid and accurate detection of lumps from breast images is a key step in assessing the condition and formulating treatment plans.

[0003] At present, the detection technology for breast lumps is mainly implemented by traditional image processing technology. Traditional image processing technologies, such as threshold segmentation, edge detection, morphological processing, etc., have been widely used in breast mass detection for a long time. These methods try to identify and mark the mass area by preprocessing the image and extracting features. However, due to the complexity of breast images, such as the diversity of tissue structure and uneven lighting, traditional image processing technology is often unable to cope with these complex features, resulting in low accuracy of mass detection. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a method, apparatus, computer equipment and storage medium for processing breast images to solve the technical problem that the traditional image processing technology used in the existing breast lump detection technology has low accuracy in lump detection.

[0005] In order to solve the above technical problems, the present application embodiment provides a method for processing breast images, which adopts the following technical solution:

[0006] receiving a breast image to be processed;

[0007] Preprocessing the breast image to obtain a corresponding target breast image;

[0008] Extracting a breast region image from the target breast image;

[0009] Calling a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module;

[0010] Performing detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result;

[0011] The mass detection result is output.

[0012] Furthermore, before the step of calling the pre-trained mass detection model, the method further includes:

[0013] Acquiring pre-collected breast image data;

[0014] Annotating the breast image data based on a preset label processing strategy to obtain corresponding image sample data;

[0015] Dividing the image sample data into a first training set and a first test set;

[0016] Calling the regional convolutional neural network, adding the feature pyramid network and the non-local module to the regional convolutional neural network, and using the focal loss function to replace the original cross entropy loss function in the regional convolutional neural network to obtain the target regional convolutional neural network;

[0017] Training the target area convolutional neural network based on the first training set to obtain a corresponding first model;

[0018] Performing model optimization on the first model based on the first test set to obtain a second model that meets preset optimization requirements;

[0019] The second model is used as the tumor detection model.

[0020] Furthermore, the step of preprocessing the breast image to obtain a corresponding target breast image specifically includes:

[0021] Performing window adjustment processing on the breast image to obtain a corresponding first image;

[0022] Performing artifact and noise removal processing on the first image to obtain a corresponding second image;

[0023] Performing enhancement processing on the second image to obtain a corresponding third image;

[0024] The third image is used as the target breast image.

[0025] Furthermore, the step of extracting a breast region image from the target breast image specifically includes:

[0026] Call the preset target detection network;

[0027] Performing breast detection processing on the target breast image based on the target detection network to obtain corresponding breast position coordinates;

[0028] Performing cropping processing on the target breast image based on the breast position coordinates to obtain a corresponding cropped image;

[0029] The cropped image is used as the breast region image.

[0030] Furthermore, the step of outputting the mass detection result specifically includes:

[0031] Performing information analysis on the mass detection result to extract key information corresponding to a preset information type;

[0032] Fill the key information into a preset test report template to obtain a corresponding target test report;

[0033] Call the preset display interface;

[0034] The target detection report is displayed in the display interface.

[0035] Further, after the step of extracting the breast region image from the target breast image, the method further includes:

[0036] Call the pre-trained calcification segmentation model;

[0037] Performing detection processing on the breast region image based on the calcification segmentation model to obtain corresponding calcification point detection results;

[0038] The calcification point detection result is output.

[0039] Furthermore, before the step of calling the pre-trained calcification segmentation model, the method further includes:

[0040] Acquire the pre-constructed image sample data, and divide the image sample data into a second training set and a second test set;

[0041] Calling a preset target U-Net network; wherein the target U-Net network is a network obtained after introducing a group normalization layer into a preset initial U-Net network;

[0042] Training the target U-Net network based on the second training set to obtain a corresponding third model;

[0043] Obtaining target evaluation indicators corresponding to the calcification segmentation task;

[0044] Based on the target evaluation index, using the second test set to perform performance evaluation on the third model to obtain a corresponding performance evaluation result;

[0045] Adjusting the third model based on the performance evaluation result until a fourth model that meets preset performance requirements is obtained;

[0046] The fourth model is used as the calcification segmentation model.

[0047] In order to solve the above technical problems, the embodiment of the present application further provides a breast image processing device, which adopts the following technical solution:

[0048] A receiving module, used for receiving a breast image to be processed;

[0049] A preprocessing module, used for preprocessing the breast image to obtain a corresponding target breast image;

[0050] An extraction module, used for extracting a breast region image from the target breast image;

[0051] A first calling module is used to call a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module;

[0052] A first detection module, configured to perform detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result;

[0053] The first output module is used to output the mass detection result.

[0054] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0055] receiving a breast image to be processed;

[0056] Preprocessing the breast image to obtain a corresponding target breast image;

[0057] Extracting a breast region image from the target breast image;

[0058] Calling a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module;

[0059] Performing detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result;

[0060] The mass detection result is output.

[0061] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0062] receiving a breast image to be processed;

[0063] Preprocessing the breast image to obtain a corresponding target breast image;

[0064] Extracting a breast region image from the target breast image;

[0065] Calling a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module;

[0066] Performing detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result;

[0067] The mass detection result is output.

[0068] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0069] The present application first receives a breast image to be processed; then preprocesses the breast image to obtain a corresponding target breast image; then extracts a breast region image from the target breast image; subsequently calls a pre-trained mass detection model; further detects and processes the breast region image based on the mass detection model to obtain a corresponding mass detection result; and finally outputs the mass detection result. The present application preprocesses the received breast image to be processed to obtain a corresponding target breast image, and extracts the breast region image from the target breast image, and then based on the pre-trained mass detection model, can quickly and accurately complete the mass detection processing of the breast region image and generate the corresponding mass detection result, effectively improving the detection accuracy of mass detection and ensuring the accuracy and reliability of the generated mass detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0071] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0072] Figure 2 A flowchart of an embodiment of a method for processing breast images according to the present application;

[0073] Figure 3 is a structural schematic diagram of an embodiment of a breast image processing device according to the present application;

[0074] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0075] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0076] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0077] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0078] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0079] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0080] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0081] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0082] It should be noted that the breast image processing method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the breast image processing device is generally arranged in the server / terminal device.

[0083] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0084] Continue to refer Figure 2 , a flow chart of an embodiment of a method for processing breast images according to the present application is shown. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted. The method for processing breast images provided in the embodiment of the present application can be applied to any scenario where breast image processing is required, and the method for processing breast images can be applied to products in these scenarios, for example, breast image detection and processing in the field of digital medicine. The method for processing breast images comprises the following steps:

[0085] Step S201, receiving a breast image to be processed.

[0086] In this embodiment, the method for processing breast images is executed on an electronic device (eg, Figure 1 The server / terminal device shown in the figure) can obtain the breast image to be processed through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future. The executor of the present application may specifically be an image processing system, or simply referred to as a system. The above-mentioned breast image may be a breast mammography X-ray image, which may also be called a breast X-ray photography image, or other images may be used, which is not limited in this embodiment. The breast image can be acquired by photographing with existing medical imaging equipment.

[0087] Step S202: preprocess the breast image to obtain a corresponding target breast image.

[0088] In this embodiment, the specific implementation process of preprocessing the breast image to obtain the corresponding target breast image will be further described in detail in subsequent specific embodiments of the present application, and will not be elaborated on here.

[0089] Step S203: extracting a breast region image from the target breast image.

[0090] In this embodiment, the specific implementation process of extracting the breast region image from the target breast image will be described in further detail in subsequent specific embodiments of the present application, and will not be elaborated on here.

[0091] Step S204, calling a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target area convolutional neural network based on pre-constructed image sample data; the target area convolutional neural network is a network obtained by optimizing a preset area convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module.

[0092] In this embodiment, the model construction process of the above-mentioned mass detection model will be further described in detail in subsequent specific embodiments of the present application, and will not be elaborated in detail here.

[0093] Step S205: performing detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result.

[0094] In this embodiment, by inputting the above breast region image into a pre-trained mass detection model, the mass detection model will perform a deep analysis on the breast region image and output a corresponding mass detection result, which includes information such as the location, size, and shape of the detected mass.

[0095] Step S206: output the mass detection result.

[0096] In this embodiment, the specific implementation process of outputting the mass detection result will be further described in detail in subsequent specific embodiments of the present application, and will not be elaborated in detail here.

[0097] The present application first receives a breast image to be processed; then preprocesses the breast image to obtain a corresponding target breast image; then extracts a breast region image from the target breast image; subsequently calls a pre-trained mass detection model; further detects and processes the breast region image based on the mass detection model to obtain a corresponding mass detection result; and finally outputs the mass detection result. The present application preprocesses the received breast image to be processed to obtain a corresponding target breast image, and extracts the breast region image from the target breast image, and then based on the pre-trained mass detection model, can quickly and accurately complete the mass detection processing of the breast region image and generate the corresponding mass detection result, effectively improving the detection accuracy of mass detection and ensuring the accuracy and reliability of the generated mass detection result.

[0098] In some optional implementations, before step S204, the electronic device may further perform the following steps:

[0099] Acquire pre-acquired breast image data.

[0100] In this embodiment, mammographic X-ray images can be obtained from a hospital's picture archiving and communication system (PACS) or other imaging sources, and the mammographic X-ray images are converted into a format suitable for model processing (such as JPEG, PNG, etc.), and the mammographic X-ray images are size normalized to ensure that the images input into the model have a uniform size, thereby obtaining processed breast image data.

[0101] The breast image data is labeled based on a preset label processing strategy to obtain corresponding image sample data.

[0102] In this embodiment, the process of labeling the breast image data based on the preset label processing strategy includes: first, using an automated tool to perform preliminary labeling on the breast image data, such as using a target detection algorithm (such as YOLO) to automatically detect the mass area. Then optimize the results of the preliminary labeling, and use semi-automatic tools to manually adjust to ensure the accuracy of the labeling. And the accuracy requirements of the labeling, such as the minimum labeling box size, the minimum labeling spacing, etc., can be set to further ensure the quality of the labeling. Then save the optimized labeling results in a specified format file and use them as corresponding image sample data for subsequent model training. Further, the labeling results will also be subjected to labeling quality inspection and labeling result correction processing. Specifically, the labeling quality inspection includes: randomly extracting a part of the samples from the labeling results for manual inspection to ensure the accuracy of the labeling. Then check whether the labeling results between different labeling tools or different labeling personnel are consistent to ensure the reliability of the labeling. And ensure that all areas that need to be labeled are correctly labeled without omissions. The labeling result correction includes: manually correcting the labeling errors found in the inspection to ensure the accuracy of the labeling. The corrected annotation results are fed back to the automated tool to optimize the algorithm and improve the accuracy of automated annotation.

[0103] By using label processing strategies to annotate breast image data, the consistency of annotations and the quality of data can be effectively improved, thereby enhancing the learning ability of the model.

[0104] The image sample data is divided into a first training set and a first test set.

[0105] In this embodiment, the image sample data can be divided into a first training set and a first test set by using a preset first division ratio. There is no limitation on the value of the first division ratio, which can be set according to actual business needs, for example, 8:2.

[0106] The regional convolutional neural network is called, and the feature pyramid network and the non-local module are added to the regional convolutional neural network, and the focal loss function is used to replace the original cross entropy loss function in the regional convolutional neural network to obtain the target regional convolutional neural network.

[0107] In this embodiment, the target area convolutional neural network is a network obtained by optimizing the preset regional convolutional neural network (Faster R-CNN) using a feature pyramid network, a focal loss function, and a non-local module. The feature pyramid network refers to the Feature Pyramid Networks (FPN) module. FPN significantly improves the model's ability to detect masses of different sizes through multi-scale feature fusion. The focal loss function refers to Focal Loss. Focal Loss solves the problem of class imbalance and is particularly suitable for situations where the ratio of positive and negative samples in mass detection is disparate. The non-local module refers to Non-Local Neural Networks, i.e., the Non-Local module. The Non-Local module enhances the model's ability to capture long-distance dependencies and improves the accuracy of detection.

[0108] Specifically, an FPN module is added on the basis of the regional convolutional neural network. The FPN module fuses feature maps of different scales to generate feature maps with multi-scale information. The feature map generated by FPN is used as the input of the RPN (Region Proposal Network) and the classifier in the subsequent regional convolutional neural network. In addition, Focal Loss is introduced into the loss function of the regional convolutional neural network. Focal Loss adjusts the weights of positive and negative samples and the weights of difficult and easy samples, so that the model pays more attention to difficult-to-classify samples and minority class samples. And during the training process, Focal Loss is used to replace the original cross entropy loss function. In addition, a Non-Local module is added to certain layers of the regional convolutional neural network (such as after certain residual blocks of ResNet). The Non-Local module generates global context information by calculating the similarity between different positions in the feature map. The output of the Non-Local module can be fused with the original feature map as the input of the subsequent layer.

[0109] The target area convolutional neural network is trained based on the first training set to obtain a corresponding first model.

[0110] In this embodiment, the target area convolutional neural network is trained by using the above-mentioned first training set, and the hyperparameters of the model are adjusted using grid search, random search and other methods, so that the model learns the mapping relationship between the breast image data and the label data contained in the first training set, and obtains the trained first model.

[0111] The first model is optimized based on the first test set to obtain a second model that meets preset optimization requirements.

[0112] In this embodiment, the trained first model is evaluated by using the first test set, the average precision (AP), recall rate and other indicators of the first model are calculated, and the deficiencies of the first model are analyzed according to the calculated evaluation results. Further adjustments and optimizations (such as adding data enhancement methods, adjusting model structures, etc.) are performed to address the deficiencies of the first model, and the training and evaluation process is repeated until the performance of the first model is satisfactory, thereby obtaining a second model that meets the optimization requirements and serves as the final mass detection model.

[0113] The second model is used as the tumor detection model.

[0114] In this embodiment, by combining a variety of advanced neural network technologies, including FPN, Focal Loss and Non-Local Neural Networks, and an optimized annotation method, the mass detection model is constructed, so that the use of the mass detection model can effectively improve the detection performance and detection accuracy of mass detection for breast X-ray images. After the mass detection model is constructed, the mass detection model can be integrated into the existing medical system, thereby improving the work efficiency of doctors and the quality of disease analysis.

[0115] The present application obtains pre-collected breast image data; and annotates the breast image data based on a preset label processing strategy to obtain corresponding image sample data; then divides the image sample data into a first training set and a first test set; then calls the regional convolutional neural network, adds the feature pyramid network and the non-local module to the regional convolutional neural network, and uses the focal loss function to replace the original cross entropy loss function in the regional convolutional neural network to obtain the target regional convolutional neural network; and trains the target regional convolutional neural network based on the first training set to obtain a corresponding first model; subsequently optimizes the first model based on the first test set to obtain a second model that meets the preset optimization requirements; finally, uses the second model as the mass detection model. The present application annotates pre-collected breast image data using a label processing strategy to obtain corresponding image sample data, and divides the image sample data into a first training set and a first test set, and then trains a target area convolutional neural network based on the first training set to obtain a corresponding first model, and then optimizes the first model using the first test set, thereby achieving rapid and accurate construction of a mass detection model that meets preset optimization requirements, effectively improving the model construction efficiency of the mass detection model, and ensuring the model effect of the obtained mass detection model.

[0116] In some optional implementations of this embodiment, step S202 includes the following steps:

[0117] A window adjustment process is performed on the breast image to obtain a corresponding first image.

[0118] In this embodiment, the above-mentioned window adjustment processing includes: using window adjustment technology (such as linear transformation) to map the image pixel value to a new range, so that the mass, calcification point, etc. in the image are more prominent. And adjusting the contrast and brightness of the image to further improve the visibility of the mass and calcification point in the image, thereby improving the image quality. Among them, the contrast of the image can be adjusted by using methods such as histogram equalization and adaptive histogram equalization. And adjusting the brightness of the image by linear transformation or nonlinear transformation to ensure that the overall brightness of the image is moderate.

[0119] Perform artifact and noise removal processing on the first image to obtain a corresponding second image.

[0120] In this embodiment, artifacts (such as marks, device shadows, etc.) in the image can be identified and removed by using morphological operations, image restoration algorithms, etc. In addition, noise in the image can be removed by using filters (such as Gaussian filters, median filters, etc.).

[0121] Perform enhancement processing on the second image to obtain a corresponding third image.

[0122] In this embodiment, the enhancement processing may include contrast stretching, sharpening, etc., to further improve the visibility of masses and calcification points in the image.

[0123] The third image is used as the target breast image.

[0124] The present application performs window adjustment processing on the breast image to obtain a corresponding first image; then performs artifact and noise removal processing on the first image to obtain a corresponding second image; then performs enhancement processing on the second image to obtain a corresponding third image; and subsequently uses the third image as the target breast image. The present application performs window adjustment processing, artifact and noise removal processing, and enhancement processing on the breast image, thereby efficiently and accurately completing the preprocessing of the breast image, ensuring the image quality of the target breast image obtained, so that the subsequent use based on the target breast image is conducive to improving the processing accuracy and detection effect of the subsequent mass detection processing of the mass detection model.

[0125] In some optional implementations, step S203 includes the following steps:

[0126] Call the preset target detection network.

[0127] In this embodiment, the target detection network may specifically adopt the YOLOv3 target detection network.

[0128] The target breast image is subjected to breast detection processing based on the target detection network to obtain corresponding breast position coordinates.

[0129] In this embodiment, the target breast image is input into a target detection network to perform breast detection to obtain corresponding breast position coordinates.

[0130] The target breast image is cropped based on the breast position coordinates to obtain a corresponding cropped image.

[0131] In this embodiment, the target breast image after preprocessing is cropped according to the obtained breast position coordinates to obtain a corresponding cropped image, thereby obtaining a final breast region image.

[0132] The cropped image is used as the breast region image.

[0133] In this embodiment, the cropped image can be further post-processed, such as removing small areas, filling holes, etc., to obtain a breast region image corresponding to a cleaner breast region. By accurately extracting the breast region image from the breast image, background interference is removed, which is beneficial to improving the robustness of the model.

[0134] The present application calls a preset target detection network; then performs breast detection processing on the target breast image based on the target detection network to obtain corresponding breast position coordinates; then performs cropping processing on the target breast image based on the breast position coordinates to obtain a corresponding cropped image; and subsequently uses the cropped image as the breast region image. The present application performs breast detection processing on the target breast image based on the use of a target detection network to obtain corresponding breast position coordinates, and then performs cropping processing on the target breast image based on the obtained breast position coordinates, thereby achieving rapid and accurate extraction of the corresponding breast region image from the target breast image, improving the extraction efficiency of the breast region image, and ensuring the image accuracy of the obtained breast region image.

[0135] In some optional implementations, step S206 includes the following steps:

[0136] The mass detection result is analyzed to extract key information corresponding to a preset information type.

[0137] In this embodiment, the above information types may at least include types such as location, size, shape, etc. Specifically, the above key information may at least include the location, size, shape of the mass, and the location, size, shape of the calcification point.

[0138] Fill the key information into the preset test report template to obtain the corresponding target test report.

[0139] In this embodiment, the test report template may be a template file constructed according to actual business needs. Exemplarily, the test report template may include a title, patient information, test date, test result description, etc. Among them, the extracted key information may be filled into the corresponding position in the preset test report template, thereby automatically generating a matching target test report.

[0140] Call the preset display interface.

[0141] In this embodiment, the above-mentioned display interface is a pre-built system front-end interface for data display.

[0142] The target detection report is displayed in the display interface.

[0143] In this embodiment, the target detection report can be displayed in an intuitive manner in the display interface, such as highlighting the detected masses or calcification points recorded in the target detection report. The highlighting and charting can be realized using an image processing library or a graphics library.

[0144] The present application performs information analysis on the mass detection result, extracts key information corresponding to the preset information type, then fills the key information into the preset detection report template to obtain the corresponding target detection report, then calls the preset display interface, and subsequently displays the target detection report in the preset display interface. The present application performs information analysis on the mass detection result, extracts key information corresponding to the preset information type, and fills the key information into the preset detection report template to obtain the corresponding target detection report, and then displays the target detection report in the preset display interface, thereby realizing the output processing of the mass detection result in an intuitive way of interface display, improving the display intelligence of the mass detection result, and helping to improve the user experience.

[0145] In some optional implementations of this embodiment, after step S203, the electronic device may further perform the following steps:

[0146] Call the pre-trained calcification segmentation model.

[0147] In this embodiment, the construction process of the calcification segmentation model will be described in further detail in subsequent specific embodiments of the present application, and will not be elaborated on here.

[0148] The breast region image is detected and processed based on the calcification segmentation model to obtain corresponding calcification point detection results.

[0149] In this embodiment, by inputting the above breast region image into a pre-trained calcification segmentation model, the calcification segmentation model will perform a deep analysis on the breast region image and output a corresponding calcification point detection result. The calcification point detection result includes information such as the position, size, and shape of the detected calcification point.

[0150] The calcification point detection result is output.

[0151] In this embodiment, the output process of outputting the above calcification point detection result can refer to the above processing process of outputting the above mass detection result, and will not be elaborated in detail here.

[0152] The present application calls a pre-trained calcification segmentation model; then detects and processes the breast region image based on the calcification segmentation model to obtain corresponding calcification point detection results; and subsequently outputs the calcification point detection results. After extracting the breast region image from the target breast image, the present application further detects and processes the breast region image based on the use of a pre-trained calcification segmentation model to obtain corresponding calcification point detection results, and outputs the calcification point detection results, effectively improving the processing efficiency of calcification segmentation of the breast region image and ensuring the accuracy of the obtained calcification point detection results.

[0153] In some optional implementations of this embodiment, before the step of calling the pre-trained calcification segmentation model, the electronic device may further perform the following steps:

[0154] The pre-constructed image sample data is obtained, and the image sample data is divided into a second training set and a second test set.

[0155] In this embodiment, the image sample data can be divided into a second training set and a second test set by using a preset second division ratio. There is no limitation on the value of the second division ratio, which can be set according to actual business needs, for example, 8:2.

[0156] Call a preset target U-Net network; wherein the target U-Net network is a network obtained after introducing a group normalization layer into a preset initial U-Net network.

[0157] In this embodiment, a group normalization layer can be added after the convolution layer of the above-mentioned initial U-Net network to replace the traditional batch normalization layer. And the number of groups for group normalization is set, for example, to a smaller integer, such as 8 or 16, and other related parameters can be adjusted. In the subsequent training process, the impact of group normalization on model performance is monitored and adjusted as needed. Among them, the construction process of the above-mentioned initial U-Net network includes: introducing new modules or layers on the basis of the U-Net network, such as attention mechanism modules, deep separable convolutions, etc., to improve the performance of the model. Then adjust the encoder-decoder structure of the U-Net network, increase or decrease the number of convolution layers to meet the needs of the calcification segmentation task. By introducing group normalization technology in the initial U-Net network, the stability and generalization ability of the model are effectively improved.

[0158] The target U-Net network is trained based on the second training set to obtain a corresponding third model.

[0159] In this embodiment, based on the above-mentioned second training set, a target U-Net network is trained using an appropriate activation function (such as a ReLU function) and a loss function (such as a cross entropy loss), and grid search, random search, or Bayesian optimization methods are used to adjust the hyperparameters of the target U-Net network (such as a learning rate, batch size, number of iterations, etc.), so that the model learns the mapping relationship between the breast image data contained in the second training set and the label data, and obtains a trained third model.

[0160] Obtain the target evaluation index corresponding to the calcification segmentation task.

[0161] In this embodiment, evaluation indicators for evaluating the performance of calcification segmentation are pre-designed for the calcification segmentation task, and may include the accuracy of calcification point detection or the completeness of calcification region segmentation.

[0162] Based on the target evaluation index, the second test set is used to perform performance evaluation on the third model to obtain a corresponding performance evaluation result.

[0163] In this embodiment, the performance of the third model can be evaluated by using the above target evaluation index on the second test set, and a corresponding performance evaluation result can be obtained.

[0164] The third model is adjusted based on the performance evaluation result until a fourth model that meets preset performance requirements is obtained.

[0165] In this embodiment, based on the obtained performance evaluation results, it is possible to analyze which types of calcification points the third model performs poorly on (such as tiny calcification points, dense calcification areas, etc.). In view of these deficiencies, the third model is further adjusted and optimized, such as adding data enhancement methods, adjusting the model structure, introducing new loss functions, etc. The training and evaluation process is repeated until the performance of the model is satisfactory, thereby obtaining the corresponding fourth model as the final calcification segmentation model.

[0166] The fourth model is used as the calcification segmentation model.

[0167] In this embodiment, by introducing a group normalization layer in the U-Net network to complete the construction of the calcification segmentation model, the stability and generalization ability of the calcification segmentation model can be improved, so that the use of the calcification segmentation model can effectively improve the detection performance and detection accuracy of calcification point detection for breast X-ray images. In addition, after the construction of the calcification segmentation model is completed, the calcification segmentation model can be integrated into the existing medical system, thereby improving the work efficiency of doctors and the quality of disease analysis.

[0168] The present application obtains the pre-constructed image sample data and divides the image sample data into a second training set and a second test set; then calls a preset target U-Net network; then trains the target U-Net network based on the second training set to obtain a corresponding third model; subsequently obtains a target evaluation index corresponding to the calcification segmentation task; and based on the target evaluation index, uses the second test set to perform a performance evaluation on the third model to obtain a corresponding performance evaluation result; further adjusts the third model based on the performance evaluation result until a fourth model that meets the preset performance requirements is obtained; finally, the fourth model is used as the calcification segmentation model. The present application divides pre-constructed image sample data into a second training set and a second test set, and then trains the target U-Net network obtained after introducing a group normalization layer in the initial U-Net network based on the second training set to obtain a corresponding third model, and based on preset target evaluation indicators, uses the second test set to perform performance evaluation on the third model to obtain corresponding performance evaluation results, and then adjusts the third model based on the performance evaluation results, so as to quickly and accurately construct a calcification segmentation model that meets preset performance requirements, effectively improve the model construction efficiency of the calcification segmentation model, and ensure the model effect of the obtained calcification segmentation model.

[0169] In some optional implementations, the user information obtained is subject to the user's consent and complies with relevant laws and policies.

[0170] In addition, the non-Company software tools or components that appear in the embodiments of the present application are merely examples and do not represent actual use.

[0171] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0172] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned tumor detection results, the above-mentioned tumor detection results can also be stored in a node of a blockchain.

[0173] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0174] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0175] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0176] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0177] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0178] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a processing device for breast images, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0179] like Figure 3As shown, the breast image processing device 300 described in this embodiment includes: a receiving module 301, a preprocessing module 302, an extraction module 303, a first calling module 304, a first detection module 305 and a first output module 306. Among them:

[0180] A receiving module 301 is used to receive a breast image to be processed;

[0181] A preprocessing module 302 is used to preprocess the breast image to obtain a corresponding target breast image;

[0182] An extraction module 303 is used to extract a breast region image from the target breast image;

[0183] The first calling module 304 is used to call a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module;

[0184] A first detection module 305 is used to perform detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result;

[0185] The first output module 306 is used to output the mass detection result.

[0186] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0187] In some optional implementations of this embodiment, the apparatus for processing breast images further includes:

[0188] A first acquisition module, used to acquire pre-collected breast image data;

[0189] A labeling module, used to label the breast image data based on a preset label processing strategy to obtain corresponding image sample data;

[0190] A first division module, used for dividing the image sample data into a first training set and a first test set;

[0191] A first calling module is used to call the regional convolutional neural network, add the feature pyramid network and the non-local module in the regional convolutional neural network, and use the focal loss function to replace the original cross entropy loss function in the regional convolutional neural network to obtain the target regional convolutional neural network;

[0192] A first training module, used to train the target area convolutional neural network based on the first training set to obtain a corresponding first model;

[0193] An optimization module, configured to optimize the first model based on the first test set to obtain a second model that meets a preset optimization requirement;

[0194] The first determination module is used to use the second model as the mass detection model.

[0195] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0196] In some optional implementations of this embodiment, the preprocessing module 302 includes:

[0197] A first processing submodule, used for performing window adjustment processing on the breast image to obtain a corresponding first image;

[0198] A second processing submodule, configured to perform artifact and noise removal processing on the first image to obtain a corresponding second image;

[0199] A third processing submodule, used for performing enhancement processing on the second image to obtain a corresponding third image;

[0200] The first determination submodule is configured to use the third image as the target breast image.

[0201] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0202] In some optional implementations of this embodiment, the extraction module 303 includes:

[0203] The first calling submodule is used to call a preset target detection network;

[0204] A detection submodule, used for performing breast detection processing on the target breast image based on the target detection network to obtain corresponding breast position coordinates;

[0205] A cropping submodule, used for cropping the target breast image based on the breast position coordinates to obtain a corresponding cropped image;

[0206] The second determination submodule is used to use the cropped image as the breast region image.

[0207] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0208] In some optional implementations of this embodiment, the first output module 306 includes:

[0209] An analysis submodule, used to analyze the mass detection result and extract key information corresponding to a preset information type;

[0210] A filling submodule is used to fill the key information into a preset detection report template to obtain a corresponding target detection report;

[0211] The second calling submodule is used to call a preset display interface;

[0212] A display submodule is used to display the target detection report in the display interface.

[0213] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0214] In some optional implementations of this embodiment, the apparatus for processing breast images further includes:

[0215] The second calling module is used to call the pre-trained calcification segmentation model;

[0216] A second detection module is used to perform detection processing on the breast region image based on the calcification segmentation model to obtain a corresponding calcification point detection result;

[0217] The second output module is used to output the calcification point detection result.

[0218] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0219] In some optional implementations of this embodiment, the apparatus for processing breast images further includes:

[0220] A second acquisition module, used for acquiring the pre-constructed image sample data, and dividing the image sample data into a second training set and a second test set;

[0221] A third calling module is used to call a preset target U-Net network; wherein the target U-Net network is a network obtained after introducing a group normalization layer into a preset initial U-Net network;

[0222] A second training module, used to train the target U-Net network based on the second training set to obtain a corresponding third model;

[0223] A fourth acquisition module is used to acquire a target evaluation index corresponding to the calcification segmentation task;

[0224] An evaluation module, configured to perform a performance evaluation on the third model using the second test set based on the target evaluation index to obtain a corresponding performance evaluation result;

[0225] An adjustment module, used for adjusting the third model based on the performance evaluation result until a fourth model that meets the preset performance requirements is obtained;

[0226] The second determining module is configured to use the fourth model as the calcification segmentation model.

[0227] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the breast image processing method in the above-mentioned embodiment, and will not be described in detail here.

[0228] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0229] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application-specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0230] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0231] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of a method for processing breast images, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0232] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for running the method for processing breast images.

[0233] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0234] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0235] In an embodiment of the present application, the present application obtains a corresponding target breast image by preprocessing the received breast image to be processed, and extracts a breast region image from the target breast image, and then based on a pre-trained mass detection model, it is possible to quickly and accurately complete the mass detection processing for the breast region image and generate a corresponding mass detection result, thereby effectively improving the detection accuracy of the mass detection and ensuring the accuracy and reliability of the generated mass detection result.

[0236] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the breast image processing method as described above.

[0237] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0238] In an embodiment of the present application, the present application obtains a corresponding target breast image by preprocessing the received breast image to be processed, and extracts a breast region image from the target breast image, and then based on a pre-trained mass detection model, it is possible to quickly and accurately complete the mass detection processing for the breast region image and generate a corresponding mass detection result, thereby effectively improving the detection accuracy of the mass detection and ensuring the accuracy and reliability of the generated mass detection result.

[0239] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0240] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for processing a breast image, characterized in that: The steps include: receiving a breast image to be processed; Preprocessing the breast image to obtain a corresponding target breast image; Extracting a breast region image from the target breast image; Calling a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module; Performing detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result; The mass detection result is output.

2. The method for processing breast images according to claim 1, characterized in that: Before the step of calling the pre-trained mass detection model, the method further includes: Acquiring pre-collected breast image data; Annotating the breast image data based on a preset label processing strategy to obtain corresponding image sample data; Dividing the image sample data into a first training set and a first test set; Calling the regional convolutional neural network, adding the feature pyramid network and the non-local module to the regional convolutional neural network, and using the focal loss function to replace the original cross entropy loss function in the regional convolutional neural network to obtain the target regional convolutional neural network; Training the target area convolutional neural network based on the first training set to obtain a corresponding first model; Performing model optimization on the first model based on the first test set to obtain a second model that meets preset optimization requirements; The second model is used as the tumor detection model.

3. The method for processing breast images according to claim 1, characterized in that: The step of preprocessing the breast image to obtain a corresponding target breast image specifically includes: Performing window adjustment processing on the breast image to obtain a corresponding first image; Performing artifact and noise removal processing on the first image to obtain a corresponding second image; Performing enhancement processing on the second image to obtain a corresponding third image; The third image is used as the target breast image.

4. The method for processing breast images according to claim 1, characterized in that: The step of extracting a breast region image from the target breast image specifically comprises: Call the preset target detection network; Performing breast detection processing on the target breast image based on the target detection network to obtain corresponding breast position coordinates; Performing cropping processing on the target breast image based on the breast position coordinates to obtain a corresponding cropped image; The cropped image is used as the breast region image.

5. The method for processing breast images according to claim 1, characterized in that: The step of outputting the mass detection result specifically includes: Performing information analysis on the mass detection result to extract key information corresponding to a preset information type; Fill the key information into a preset test report template to obtain a corresponding target test report; Call the preset display interface; The target detection report is displayed in the display interface.

6. The method for processing breast images according to claim 1, characterized in that: After the step of extracting the breast region image from the target breast image, the method further includes: Call the pre-trained calcification segmentation model; Performing detection processing on the breast region image based on the calcification segmentation model to obtain corresponding calcification point detection results; The calcification point detection result is output.

7. The method for processing breast images according to claim 6, characterized in that: Before the step of calling the pre-trained calcification segmentation model, the method further includes: Acquire the pre-constructed image sample data, and divide the image sample data into a second training set and a second test set; Calling a preset target U-Net network; wherein the target U-Net network is a network obtained after introducing a group normalization layer into a preset initial U-Net network; Training the target U-Net network based on the second training set to obtain a corresponding third model; Obtaining target evaluation indicators corresponding to the calcification segmentation task; Based on the target evaluation index, using the second test set to perform performance evaluation on the third model to obtain a corresponding performance evaluation result; Adjusting the third model based on the performance evaluation result until a fourth model that meets preset performance requirements is obtained; The fourth model is used as the calcification segmentation model.

8. A mammary gland image processing device, characterized in that: include: A receiving module, used for receiving a breast image to be processed; A preprocessing module, used for preprocessing the breast image to obtain a corresponding target breast image; An extraction module, used for extracting a breast region image from the target breast image; A first calling module is used to call a pre-trained mass detection model; wherein the mass detection model is a model obtained by training a target region convolutional neural network based on pre-constructed image sample data; the target region convolutional neural network is a network obtained by optimizing a preset region convolutional neural network using a feature pyramid network, a focal loss function, and a non-local module; A first detection module, configured to perform detection processing on the breast region image based on the mass detection model to obtain a corresponding mass detection result; The first output module is used to output the mass detection result.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the method for processing breast images according to any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the method for processing breast images according to any one of claims 1 to 7 are implemented.

Images