Multi-parameter Breast MRI Image Segmentation Method Based on Dynamic Adaptive Network

Abstract

The embodiment of the present invention provides a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network, which is applied to the field of medical imaging technology. The method includes: obtaining a multi-parameter magnetic resonance sample image; inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, training the image segmentation model to be trained to obtain a trained image segmentation model; obtaining a single-parameter magnetic resonance sample image; inputting the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtaining a trained image segmentation model; obtaining a magnetic resonance image to be recognized; and inputting the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion area in the magnetic resonance image to be recognized. Thereby, the requirements for sample images in the training process of the image segmentation model are reduced.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and particularly to a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network. Background Art

[0002] At present, with the rapid development of network model technology, network models have penetrated into many fields of people's lives. For example, face recognition, image classification, data prediction, etc. are carried out through network models. Through network models, not only can great convenience be provided to people's lives, but also the efficiency of data processing can be improved.

[0003] However, in the current field of medical imaging technology, when training a network model for identifying lesion sites in images taken by patients, it is often through multi-parameter images for model training and testing, resulting in relatively high requirements for sample images. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network to solve the problem that only multi-parameter images can be used for model training and testing. The specific technical solutions are as follows:

[0005] In the first aspect of the embodiments of the present application, first, a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network is provided, including:

[0006] Obtain multi-parameter magnetic resonance sample images, where the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion regions;

[0007] Input the multi-parameter magnetic resonance sample images into an image segmentation model to be trained, and train the image segmentation model to be trained to obtain a trained image segmentation model, where the image segmentation model to be trained is a dynamic adaptive network model;

[0008] Obtain single-parameter magnetic resonance sample images;

[0009] Input the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model. When the test is successful, a trained image segmentation model is obtained;

[0010] Obtain a magnetic resonance image to be recognized;

[0011] Input the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized.

[0012] Optionally, the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model includes:

[0013] Input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and identify the multi-parameter magnetic resonance sample image through the dynamic convolution in the image segmentation model to be trained, so as to obtain the current recognition result;

[0014] Calculate the current training loss of the image segmentation model to be trained according to the recognition result;

[0015] Adjust the parameters of the image segmentation model to be trained according to the current training loss, and return to the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through the dynamic convolution in the image segmentation model to be trained to obtain the current recognition result, and continue to execute until the preset number of iterations is reached, and a trained image segmentation model is obtained.

[0016] Optionally, the step of inputting the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and obtaining a trained image segmentation model when the test is successful includes:

[0017] Input the single-parameter magnetic resonance sample image into the trained image segmentation model to obtain the current test loss of the trained image segmentation model;

[0018] When the current test loss is less than the preset threshold, a trained image segmentation model is obtained.

[0019] Optionally, after the step of inputting the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and obtaining a trained image segmentation model when the test is successful, the method further includes:

[0020] When the current loss is greater than the preset threshold, return to the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model, and continue training.

[0021] In the second aspect of the embodiments of the present application, a multi-parameter breast magnetic resonance image segmentation device based on a dynamic adaptive network is provided, including:

[0022] A training image acquisition module, configured to acquire a multi-parameter magnetic resonance sample image, where the multi-parameter magnetic resonance sample image includes a magnetic resonance sample image containing a breast lesion area;

[0023] A model training module, configured to input the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, train the image segmentation model to be trained, and obtain a trained image segmentation model, wherein the image segmentation model to be trained is a dynamic adaptive network model;

[0024] A test image acquisition module, configured to acquire a single-parameter magnetic resonance sample image;

[0025] A model testing module, configured to input the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtain a trained image segmentation model;

[0026] A to-be-recognized image acquisition module, configured to acquire a magnetic resonance image to be recognized;

[0027] A to-be-recognized image recognition module, configured to input the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion area in the magnetic resonance image to be recognized.

[0028] Optionally, the model training module includes:

[0029] A recognition result acquisition sub-module, configured to input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and recognize the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current recognition result;

[0030] A training loss calculation sub-module, configured to calculate the current training loss of the image segmentation model to be trained according to the recognition result;

[0031] A model parameter adjustment sub-module, configured to adjust the parameters of the image segmentation model to be trained according to the current training loss, and return to execute the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and recognizing the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current recognition result, until a preset number of iterations is reached, and a trained image segmentation model is obtained.

[0032] Optionally, the model testing module includes:

[0033] A test loss calculation sub-module, configured to input the single-parameter magnetic resonance sample image into the trained image segmentation model to obtain the current test loss of the trained image segmentation model;

[0034] A model output sub-module, configured to obtain a trained image segmentation model when the current test loss is less than a preset threshold.

[0035] Optionally, the device further includes:

[0036] A continue training module, configured to, when the current loss is greater than a preset threshold, return the step of inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model, and continue the training.

[0037] On the other hand, an embodiment of the present application further provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete mutual communication through the communication bus;

[0038] The memory is used to store a computer program;

[0039] The processor is configured to, when executing the program stored on the memory, implement any one of the above-mentioned multi-parameter breast magnetic resonance image segmentation methods based on a dynamic adaptive network.

[0040] On the other hand, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any one of the above-mentioned multi-parameter breast magnetic resonance image segmentation methods based on a dynamic adaptive network is implemented.

[0041] An embodiment of the present invention further provides a computer program product including instructions, which, when running on a computer, cause the computer to execute any one of the above-mentioned multi-parameter breast magnetic resonance image segmentation methods based on a dynamic adaptive network.

[0042] Advantages of the embodiments of the present invention:

[0043] The multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network provided by an embodiment of the present invention includes obtaining multi-parameter magnetic resonance sample images, where the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion regions; inputting the multi-parameter magnetic resonance sample images into an image segmentation model to be trained, training the image segmentation model to be trained to obtain a trained image segmentation model, where the image segmentation model to be trained is a dynamic adaptive network model; obtaining single-parameter magnetic resonance sample images; inputting the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtaining a trained image segmentation model; obtaining a magnetic resonance image to be recognized; and inputting the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized. The image segmentation model to be trained can be trained using multi-parameter magnetic resonance sample images and the trained model can be tested using single-parameter magnetic resonance sample images, thereby solving the problem that only multi-parameter images can be used for model training and testing, and reducing the requirements for sample images in the training process of the image segmentation model.

[0044] Of course, when implementing any product or method of the present invention, it is not necessarily required to achieve all the above advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other embodiments based on these drawings.

[0046] Figure 1 It is a schematic flowchart of a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network provided by an embodiment of the present application;

[0047] Figure 2 It is a schematic flowchart of training an image segmentation model to be trained provided by an embodiment of the present application;

[0048] Figure 3 It is a schematic flowchart of testing a trained image segmentation model provided by an embodiment of the present application;

[0049] Figure 4 It is a schematic structural diagram of a multi-parameter breast magnetic resonance image segmentation device based on a dynamic adaptive network provided by an embodiment of the present application;

[0050] Figure 5 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0051] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art based on this application belong to the protection scope of the present invention.

[0052] In the first aspect of the embodiments of the present application, first, a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network is provided, including:

[0053] Obtain multi-parameter magnetic resonance sample images, where the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion regions;

[0054] Input the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, and train the image segmentation model to be trained to obtain a trained image segmentation model, where the image segmentation model to be trained is a dynamic adaptive network model;

[0055] Obtain single-parameter magnetic resonance sample images;

[0056] Input the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model. When the test is successful, a trained image segmentation model is obtained;

[0057] Obtain the magnetic resonance image to be recognized;

[0058] Input the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized.

[0059] It can be seen that through the method of the embodiments of the present application, the image segmentation model to be trained can be trained by multi-parameter magnetic resonance sample images, and the trained model can be tested by single-parameter magnetic resonance sample images, thereby solving the problem that only multi-parameter images can be used for training and testing of the model, and reducing the requirements for sample images in the training process of the image segmentation model.

[0060] Specifically, referring to Figure 1 , Figure 1 is a schematic flowchart of a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network provided by the embodiments of the present application, including:

[0061] Step S11: Obtain multi-parameter magnetic resonance sample images.

[0062] Among them, the multi-parameter magnetic resonance sample image includes a magnetic resonance sample image containing a breast lesion area. The multi-parameter magnetic resonance sample image in the embodiments of the present application can be a pre-acquired magnetic resonance image. Specifically, the multi-parameter magnetic resonance sample image and the single-parameter magnetic resonance sample image in the present application can refer to the prior art and will not be elaborated here.

[0063] The method of the embodiments of the present application is applied to a server or an intelligent terminal and can be implemented by the server or the intelligent terminal. Specifically, the intelligent terminal can be a computer. In actual use, the image segmentation model provided by the embodiments of the present application can be trained and tested through the server or the intelligent terminal to obtain a trained image segmentation model.

[0064] Step S12: Input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and train the image segmentation model to be trained to obtain a trained image segmentation model.

[0065] Among them, the image segmentation model to be trained is a dynamic adaptive network model. The image segmentation model to be trained provided by the embodiments of the present application can be a dynamic adaptive network. Through the image segmentation model to be trained, the rich information provided by the multi-parameter magnetic resonance imaging in the training data can be maximally absorbed for model training to obtain a trained image segmentation model.

[0066] Step S13: Obtain a single-parameter magnetic resonance sample image.

[0067] The image segmentation model to be trained provided by the embodiments of the present application can be trained through the multi-parameter magnetic resonance sample image, and then the trained model can be tested through the single-parameter magnetic resonance sample image. Specifically, the single-parameter magnetic resonance sample image in the embodiments of the present application can include a magnetic resonance sample image containing a breast lesion area.

[0068] Step S14: Input the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model. When the test is successful, a trained image segmentation model is obtained.

[0069] Among them, the method of the embodiment of the present application can be applied to the recognition of breast lesion regions. The existing breast lesion region segmentation technology based on multi-parametric magnetic resonance images often has difficulties in fusing multi-parametric image information and requires the design of special network modules. However, the generality of special fusion modules is generally poor, and the existing technology still needs to use registered multi-parametric magnetic resonance images during testing, resulting in high testing costs. The dynamic adaptive network provided by the method of the embodiment of the present application can absorb rich information of multi-parametric magnetic resonance images through adaptive adjustment of parameters without relying on special modules. At the same time, during testing, accurate segmentation can be achieved with only single-parametric image input, reducing the testing cost. Specifically, the structure of the adaptive grid can refer to the prior art.

[0070] Step S15, obtain the magnetic resonance image to be recognized.

[0071] The magnetic resonance image to be recognized in the embodiment of the present application can be the magnetic resonance image of a patient obtained during actual use.

[0072] Step S16, input the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized.

[0073] By inputting the magnetic resonance image to be recognized into the above-mentioned trained image segmentation model, the breast lesion region in the magnetic resonance image of the patient can be recognized, so that it can be determined whether the patient has a breast lesion and the location of the breast lesion region based on the recognition result.

[0074] It can be seen that through the method of the embodiment of the present application, the image segmentation model to be trained can be trained with multi-parametric magnetic resonance sample images, and the trained model can be tested with single-parametric magnetic resonance sample images, thus solving the problem that the model can only be trained and tested with multi-parametric images and reducing the requirements for sample images in the training process of the image segmentation model.

[0075] Optionally, referring to Figure 2 , step S12 inputs the multi-parametric magnetic resonance sample image into the image segmentation model to be trained to train the image segmentation model to be trained and obtain the trained image segmentation model, including:

[0076] Step S121, input the multi-parametric magnetic resonance sample image into the image segmentation model to be trained, and use the dynamic convolution in the image segmentation model to be trained to recognize the multi-parametric magnetic resonance sample image to obtain the current recognition result;

[0077] Step S122, calculate the current training loss of the image segmentation model to be trained according to the recognition result;

[0078] Step S123: Adjust the parameters of the image segmentation model to be trained according to the current training loss, and return to continue the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and performing recognition on the multi-parameter magnetic resonance sample image through the dynamic convolution in the image segmentation model to be trained to obtain the current recognition result, until the preset number of iterations is reached, and a trained image segmentation model is obtained.

[0079] For the dynamic adaptive network provided in the embodiments of the present application, in view of the limited conditions for the input of multi-parameter magnetic resonance images, by converting ordinary convolution operations into dynamic convolutions composed of multiple sub-convolutions, different sub-convolutions can be trained through the network to extract different image features for magnetic resonance imaging with different parameters. Moreover, since there is no pixel-level fusion of image information at any level of the image and features, multi-parameter magnetic resonance images do not need to be registered.

[0080] Among them, when adjusting the parameters of the image segmentation model to be trained according to the current training loss, the constructed multi-parameter breast magnetic resonance images can be used to optimize the overall network and learn the mapping relationship from image input to breast lesion area segmentation end-to-end. By introducing adaptive weights, different sub-convolutions can be made to focus on the extraction of information at different levels, improving the feature extraction efficiency of the model and extracting more information.

[0081] Optionally, see Figure 3 , step S14 inputs the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model. When the test is successful, a trained image segmentation model is obtained, including:

[0082] Step S141: Input the single-parameter magnetic resonance sample image into the trained image segmentation model to obtain the current test loss of the trained image segmentation model;

[0083] Step S142: When the current test loss is less than the preset threshold, a trained image segmentation model is obtained.

[0084] Optionally, after inputting the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model and obtaining a trained image segmentation model when the test is successful, the method further includes: when the current loss is greater than the preset threshold, return to the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained and training the image segmentation model to be trained to obtain a trained image segmentation model, and continue training.

[0085] Through the method of the embodiments of the present application, when testing the trained model, single-parameter magnetic resonance sample images can be used, so that in the case of only single-parameter breast magnetic resonance images input, the influence of each sub-convolution on the final result can be adaptively adjusted, so that the input can be processed most effectively, and accurate segmentation of the breast lesion area can be achieved.

[0086] For the segmentation of multi-parameter magnetic resonance images in the prior art, special network modules need to be designed to achieve the extraction and fusion of information provided by different imaging parameters. The design of special modules is relatively complex, and the generality between different data sets is generally poor. In addition, such methods require the provision of registered multi-parameter magnetic resonance images during testing, and the testing cost is relatively high. The method of the embodiments of the present application can achieve dynamic adaptive extraction of multi-parameter magnetic resonance image information, and multi-parameter images are not required during testing, while reducing the testing cost and ensuring the accuracy of segmentation.

[0087] In a second aspect of the embodiments of the present application, a multi-parameter breast magnetic resonance image segmentation device based on a dynamic adaptive network is provided. Refer to Figure 4 , including:

[0088] A training image acquisition module 401, configured to acquire multi-parameter magnetic resonance sample images, where the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion areas;

[0089] A model training module 402, configured to input the multi-parameter magnetic resonance sample images into an image segmentation model to be trained, and train the image segmentation model to be trained to obtain a trained image segmentation model, where the image segmentation model to be trained is a dynamic adaptive network model;

[0090] A test image acquisition module 403, configured to acquire single-parameter magnetic resonance sample images;

[0091] A model test module 404, configured to input the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtain a trained image segmentation model;

[0092] A to-be-recognized image acquisition module 405, configured to acquire a magnetic resonance image to be recognized;

[0093] A to-be-recognized image recognition module 406, configured to input the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion area in the magnetic resonance image to be recognized.

[0094] Optionally, the model training module 402 includes:

[0095] An identification result acquisition sub-module, configured to input a multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and perform identification on the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, so as to obtain the current identification result;

[0096] A training loss calculation sub-module, configured to calculate the current training loss of the image segmentation model to be trained according to the identification result;

[0097] A model parameter adjustment sub-module, configured to adjust the parameters of the image segmentation model to be trained according to the current training loss, and return to execute the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, and performing identification on the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain the current identification result, until a preset number of iterations is reached, so as to obtain the trained image segmentation model.

[0098] Optionally, the model testing module 404 includes:

[0099] A test loss calculation sub-module, configured to input a single-parameter magnetic resonance sample image into the trained image segmentation model to obtain the current test loss of the trained image segmentation model;

[0100] A model output sub-module, configured to obtain the trained image segmentation model when the current test loss is less than a preset threshold.

[0101] Optionally, the above device further includes:

[0102] A continuous training module, configured to, when the current loss is greater than a preset threshold, return to execute the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, training the image segmentation model to be trained, so as to obtain the trained image segmentation model, and continue training.

[0103] It can be seen that through the device of the embodiment of the present application, the image segmentation model to be trained can be trained through a multi-parameter magnetic resonance sample image, and the trained model can be tested through a single-parameter magnetic resonance sample image, thereby solving the problem that only multi-parameter images can be used for training and testing the model, and reducing the requirements for sample images in the training process of the image segmentation model.

[0104] An embodiment of the present invention further provides an electronic device, as Figure 5 shown, including a processor 501, a communication interface 502, a memory 503, and a communication bus 504, wherein the processor 501, the communication interface 502, and the memory 503 communicate with each other through the communication bus 504,

[0105] The memory 503 is used for storing a computer program;

[0106] The processor 501, when executing the program stored in the memory 503, implements the following steps:

[0107] Obtain a multi-parameter magnetic resonance sample image;

[0108] Input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, train the image segmentation model to be trained, and obtain a trained image segmentation model;

[0109] Obtain a single-parameter magnetic resonance sample image, and input the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and obtain a well-trained image segmentation model.

[0110] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity, only a thick line is used in the figure to represent it, but it does not mean that there is only one bus or one type of bus.

[0111] The communication interface is used for communication between the above electronic device and other devices.

[0112] The memory may include a Random Access Memory (RAM), or may also include a Non-Volatile Memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located far from the aforementioned processor.

[0113] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0114] In yet another embodiment provided by the present invention, there is also provided a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the above-mentioned multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network is implemented.

[0115] In yet another embodiment provided by the present invention, there is also provided a computer program product containing instructions, and when it runs on a computer, it causes the computer to execute any one of the above-mentioned multi-parameter breast magnetic resonance image segmentation methods based on a dynamic adaptive network.

[0116] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

[0117] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements includes not only those elements but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article, or device including the element.

[0118] Each embodiment in this specification is described in a related manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the embodiments of the device, electronic device, storage medium, and computer program product, since they are basically similar to the method embodiments, the description is relatively simple, and for the relevant parts, reference can be made to the partial description of the method embodiments.

[0119] The above are only the preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are all included in the protection scope of the present invention.

Claims

1. A multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network, characterized in that, it includes: Obtain multi-parameter magnetic resonance sample images, wherein the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion regions; Input the multi-parameter magnetic resonance sample images into an image segmentation model to be trained, and train the image segmentation model to be trained to obtain a trained image segmentation model, wherein the image segmentation model to be trained is a dynamic adaptive network model; Obtain single-parameter magnetic resonance sample images; Input the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model. When the test is successful, obtain a well-trained image segmentation model; Obtain the magnetic resonance image to be recognized; Input the magnetic resonance image to be recognized into the well-trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized.

2. The method according to claim 1, characterized in that, The step of inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model includes: Input the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, and use the dynamic convolution in the image segmentation model to be trained to recognize the multi-parameter magnetic resonance sample images to obtain the current recognition result; Calculate the current training loss of the image segmentation model to be trained according to the recognition result; Adjust the parameters of the image segmentation model to be trained according to the current training loss, and return to the step of inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, and using the dynamic convolution in the image segmentation model to be trained to recognize the multi-parameter magnetic resonance sample images to obtain the current recognition result, and continue to execute until the preset number of iterations is reached to obtain a trained image segmentation model.

3. The method according to claim 1, characterized in that, The step of inputting the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model. When the test is successful, obtaining a well-trained image segmentation model includes: Input the single-parameter magnetic resonance sample images into the trained image segmentation model to obtain the current test loss of the trained image segmentation model; When the current test loss is less than the preset threshold, obtain a well-trained image segmentation model.

4. The method according to claim 3, characterized in that, After the step of inputting the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtaining a well-trained image segmentation model, the method further includes: When the current loss is greater than the preset threshold, return to the step of inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model to continue training.

5. A multi-parameter breast magnetic resonance image segmentation device based on a dynamic adaptive network, characterized in that, it includes: A training image acquisition module for acquiring multi-parameter magnetic resonance sample images, wherein the multi-parameter magnetic resonance sample images include magnetic resonance sample images containing breast lesion regions; A model training module for inputting the multi-parameter magnetic resonance sample images into an image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model, wherein the image segmentation model to be trained is a dynamic adaptive network model; A test image acquisition module for acquiring single-parameter magnetic resonance sample images; A model test module for inputting the single-parameter magnetic resonance sample images into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, obtaining a trained image segmentation model; An image to be recognized acquisition module for acquiring a magnetic resonance image to be recognized; An image to be recognized recognition module for inputting the magnetic resonance image to be recognized into the trained image segmentation model to obtain information on the breast lesion region in the magnetic resonance image to be recognized.

6. The device according to claim 5, characterized in that, the model training module includes: A recognition result acquisition sub-module for inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, and recognizing the multi-parameter magnetic resonance sample images through dynamic convolution in the image segmentation model to be trained to obtain the current recognition result; A training loss calculation sub-module for calculating the current training loss of the image segmentation model to be trained according to the recognition result; A model parameter adjustment sub-module for adjusting the parameters of the image segmentation model to be trained according to the current training loss, and returning to the step of inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, and recognizing the multi-parameter magnetic resonance sample images through dynamic convolution in the image segmentation model to be trained to obtain the current recognition result, and continuing to execute until a preset number of iterations is reached to obtain a trained image segmentation model.

7. The device according to claim 5, characterized in that, the model test module includes: A test loss calculation sub-module for inputting the single-parameter magnetic resonance sample images into the trained image segmentation model to obtain the current test loss of the trained image segmentation model; A model output sub-module for obtaining a trained image segmentation model when the current test loss is less than a preset threshold.

8. The device according to claim 7, characterized in that, the device further includes: A continued training module for, when the current loss is greater than a preset threshold, returning to the step of inputting the multi-parameter magnetic resonance sample images into the image segmentation model to be trained, training the image segmentation model to be trained, and obtaining a trained image segmentation model to continue training.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus; The memory is used to store computer programs; The processor is used to implement the method steps described in any one of claims 1-5 when executing the programs stored on the memory.

10. A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method steps described in any one of claims 1-5.

Images