CN114463345A - Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network - Google Patents
Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network Download PDFInfo
- Publication number
- CN114463345A CN114463345A CN202111566049.XA CN202111566049A CN114463345A CN 114463345 A CN114463345 A CN 114463345A CN 202111566049 A CN202111566049 A CN 202111566049A CN 114463345 A CN114463345 A CN 114463345A
- Authority
- CN
- China
- Prior art keywords
- trained
- segmentation model
- image
- image segmentation
- magnetic resonance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30068—Mammography; Breast
Abstract
The embodiment of the invention provides a dynamic adaptive network-based multi-parameter mammary gland magnetic resonance image segmentation method, which is applied to the technical field of medical imaging and comprises the steps of obtaining a multi-parameter magnetic resonance sample image; inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training the image segmentation model to be trained to obtain a trained image segmentation model; acquiring a single-parameter magnetic resonance sample image; inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and obtaining the trained image segmentation model when the test is successful; acquiring a magnetic resonance image to be identified; and inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified. Therefore, the requirement on the sample image in the training process of the image segmentation model is reduced.
Description
Technical Field
The invention relates to the technical field of medical imaging, in particular to a multi-parameter mammary gland magnetic resonance image segmentation method based on a dynamic self-adaptive network.
Background
At present, with the rapid development of network model technology, network models have been deeply developed into multiple fields in people's lives. For example, face recognition, image classification, data prediction, etc. are performed by a network model. The network model not only can provide great convenience for life of people, but also can improve the efficiency of data processing.
However, in the field of medical imaging technology, when training a network model for identifying a lesion site in an image captured by a patient, training and testing of the model are often performed through a multi-parameter image, which results in higher requirements for a sample image.
Disclosure of Invention
The embodiment of the invention aims to provide a multi-parameter mammary gland magnetic resonance image segmentation method based on a dynamic self-adaptive network, which is used for solving the problem that a model can be trained and tested only through a multi-parameter image. The specific technical scheme is as follows:
in a first aspect of the embodiments of the present application, a method for segmenting a multi-parameter breast magnetic resonance image based on a dynamic adaptive network is provided, including:
acquiring a multi-parameter magnetic resonance sample image, wherein the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region;
inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training 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 self-adaptive network model;
acquiring a single-parameter magnetic resonance sample image;
inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and obtaining the trained image segmentation model when the test is successful;
acquiring a magnetic resonance image to be identified;
and inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified.
Optionally, the 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 includes:
inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and adjusting parameters of the image segmentation model to be trained according to the current training loss, returning to the image segmentation model to be trained, inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
Optionally, the 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 the trained image segmentation model includes:
inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to obtain the current test loss of the trained image segmentation model;
and when the current test loss is less than a preset threshold value, obtaining a trained image segmentation model.
Optionally, the single-parameter magnetic resonance sample image is input into a trained image segmentation model to test the trained image segmentation model, and when the test is successful, after the trained image segmentation model is obtained, the method further includes:
and when the current loss is larger than a preset threshold value, returning 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 continuing to train the step of obtaining the trained image segmentation model.
In a second aspect of the embodiments of the present application, a multi-parameter breast magnetic resonance image segmentation apparatus based on a dynamic adaptive network is provided, including:
the system comprises a training image acquisition module, a data acquisition module and a data processing module, wherein the training image acquisition module is used for acquiring a multi-parameter magnetic resonance sample image, and the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region;
the model training module is used for 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 the trained image segmentation model, wherein the image segmentation model to be trained is a dynamic self-adaptive network model;
the test image acquisition module is used for acquiring a single-parameter magnetic resonance sample image;
the model testing module is used for inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and when the test is successful, the trained image segmentation model is obtained;
the to-be-identified image acquisition module is used for acquiring a to-be-identified magnetic resonance image;
and the to-be-identified image identification module is used for inputting the to-be-identified magnetic resonance image into the trained image segmentation model to obtain the information of the breast lesion area in the to-be-identified magnetic resonance image.
Optionally, the model training module includes:
the identification result acquisition submodule is used for inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
the training loss calculation submodule is used for calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and the model parameter adjusting submodule is used for adjusting parameters of the image segmentation model to be trained according to the current training loss, returning the multi-parameter magnetic resonance sample image to the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
Optionally, the model testing module includes:
the test loss calculation submodule is used for inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to obtain the current test loss of the trained image segmentation model;
and the model output submodule is used for obtaining a trained image segmentation model when the current test loss is less than a preset threshold value.
Optionally, the apparatus further comprises:
and the continuous training module is used for returning to the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained when the current loss is greater than a preset threshold value, training the image segmentation model to be trained, and continuously training the step of obtaining the trained image segmentation model.
On the other hand, the embodiment of the present application further provides an electronic device, which includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing any one of the above multi-parameter mammary gland magnetic resonance image segmentation methods based on the dynamic self-adaptive network when executing the program stored in the memory.
In another aspect of the embodiments of the present application, a computer-readable storage medium is further provided, in which a computer program is stored, and the computer program, when executed by a processor, implements any one of the above dynamic adaptive network-based multi-parameter breast magnetic resonance image segmentation methods.
Embodiments of the present invention further provide a computer program product containing instructions, which when run on a computer, causes the computer to perform any one of the above-mentioned methods for dynamic adaptive network-based multi-parameter breast magnetic resonance image segmentation.
The embodiment of the invention has the following beneficial effects:
the embodiment of the invention provides a dynamic adaptive network-based multi-parameter mammary gland magnetic resonance image segmentation method, which comprises the steps of obtaining a multi-parameter magnetic resonance sample image, wherein the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a mammary gland lesion area; inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training 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 self-adaptive network model; acquiring a single-parameter magnetic resonance sample image; inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and obtaining the trained image segmentation model when the test is successful; acquiring a magnetic resonance image to be identified; and inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified. The image segmentation model to be trained can be trained through the multi-parameter magnetic resonance sample image, and the trained model is tested through the single-parameter magnetic resonance sample image, so that the problem that the model can only be trained and tested through the multi-parameter image is solved, and the requirement on the sample image in the training process of the image segmentation model is reduced.
Of course, not all of the advantages described above need to be achieved at the same time in the practice of any one product or method of the invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other embodiments can be obtained by referring to these drawings.
Fig. 1 is a schematic flowchart of a multi-parameter breast magnetic resonance image segmentation method based on a dynamic adaptive network according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a process of training an image segmentation model to be trained according to an embodiment of the present application;
fig. 3 is a schematic flowchart of testing a trained image segmentation model according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a multi-parameter breast magnetic resonance image segmentation apparatus based on a dynamic adaptive network according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived from the embodiments given herein by one of ordinary skill in the art, are within the scope of the invention.
In a first aspect of the embodiments of the present application, a method for segmenting a multi-parameter breast magnetic resonance image based on a dynamic adaptive network is provided, including:
acquiring a multi-parameter magnetic resonance sample image, wherein the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region;
inputting a multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training 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 self-adaptive network model;
acquiring 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 obtaining the trained image segmentation model when the test is successful;
acquiring a magnetic resonance image to be identified;
and inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified.
Therefore, by the method of the embodiment of the application, the image segmentation model to be trained can be trained through the multi-parameter magnetic resonance sample image, and the trained model is tested through the single-parameter magnetic resonance sample image, so that the problem that the model can only be trained and tested through the multi-parameter image is solved, and the requirement on the sample image in the training process of the image segmentation model is reduced.
Specifically, referring to fig. 1, fig. 1 is a schematic flow chart of a method for segmenting a multi-parameter breast magnetic resonance image based on a dynamic adaptive network according to an embodiment of the present application, including:
in step S11, a multi-parameter magnetic resonance sample image is acquired.
Wherein the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region. The multi-parameter magnetic resonance sample image in the embodiments of the present application may 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 be referred to in the prior art, and are not described herein again.
The method of the embodiment of the application is applied to a server or an intelligent terminal, and can be implemented through the server or the intelligent terminal, and specifically, the intelligent terminal can be a computer. In an actual use process, the image segmentation model provided by the embodiment of the application can be trained and tested through the server or the intelligent terminal, so that the trained image segmentation model is obtained.
And step S12, 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 the trained image segmentation model.
The image segmentation model to be trained is a dynamic self-adaptive network model. The image segmentation model to be trained provided by the embodiment of the application can be a dynamic self-adaptive network. The image segmentation model to be trained can absorb rich information provided by multi-parameter magnetic resonance imaging in training data to the maximum extent and train the model to obtain the trained image segmentation model.
In step S13, a single parameter magnetic resonance sample image is acquired.
The image segmentation model to be trained provided by the embodiment of the application can be used for training the model through the multi-parameter magnetic resonance sample image, and then testing the trained model through the single-parameter magnetic resonance sample image. In particular, the single-parameter magnetic resonance sample image in the embodiment of the present application may include a magnetic resonance sample image including a breast lesion region.
And step S14, inputting the single-parameter magnetic resonance sample image into the trained image segmentation model to test the trained image segmentation model, and obtaining the trained image segmentation model when the test is successful.
The method provided by the embodiment of the application can be applied to identification of the breast lesion area, the existing breast lesion area segmentation technology based on the multi-parameter magnetic resonance image is difficult to fuse multi-parameter image information, a special network module needs to be designed, the universality of the special fusion module is generally poor, the registered multi-parameter magnetic resonance image still needs to be used in the testing process of the prior art, and the testing cost is high. The dynamic self-adaptive network provided by the method of the embodiment of the application can absorb rich information of the multi-parameter magnetic resonance image through self-adaptive adjustment of the parameters without depending on a special module. Meanwhile, during testing, accurate segmentation can be realized only by inputting a single-parameter image, and the testing cost is reduced. In particular, the structure of the adaptive mesh can be seen in the prior art.
In step S15, a magnetic resonance image to be identified is acquired.
The magnetic resonance image to be identified in the embodiment of the present application may be a magnetic resonance image of a patient acquired during actual use.
And step S16, inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified.
The breast lesion area in the magnetic resonance image of the patient can be identified by inputting the magnetic resonance image to be identified into the trained image segmentation model, so that whether the patient suffers from breast lesions or not and the position of the breast lesion area can be judged according to the identification result.
Therefore, by the method of the embodiment of the application, the image segmentation model to be trained can be trained through the multi-parameter magnetic resonance sample image, and the trained model is tested through the single-parameter magnetic resonance sample image, so that the problem that the model can only be trained and tested through the multi-parameter image is solved, and the requirement on the sample image in the training process of the image segmentation model is reduced.
Optionally, referring to fig. 2, in step S12, 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, where the method includes:
step S121, inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
step S122, calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and S123, adjusting parameters of the image segmentation model to be trained according to the current training loss, returning to input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
According to the dynamic self-adaptive network provided by the embodiment of the application, aiming at the limited condition of multi-parameter magnetic resonance image input, common convolution operation is converted into dynamic convolution formed by combining a plurality of sub-convolution products, and different sub-convolution products can extract different image characteristics aiming at magnetic resonance imaging of different parameters through network training. Moreover, since there is no fusion of image information pixel levels of any layer of image and feature, the multi-parameter magnetic resonance image does not need to be registered.
When parameters of an image segmentation model to be trained are adjusted according to the current training loss, the constructed multi-parameter breast magnetic resonance image can be used to optimize the whole network, and the mapping relation of the image input to the breast lesion region segmentation is learned end to end. By introducing the self-adaptive weight, different sub-convolution focuses on the extraction of information of different levels, the feature extraction efficiency of the model is improved, and more information is extracted.
Optionally, referring to fig. 3, in step S14, 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 the trained image segmentation model, including:
step S141, inputting 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;
and step S142, when the current test loss is less than a preset threshold value, obtaining a trained image segmentation model.
Optionally, the method further includes, after the single-parameter magnetic resonance sample image is input into the trained image segmentation model to test the trained image segmentation model, and when the test is successful, the trained image segmentation model is obtained: and when the current loss is larger than the preset threshold value, returning to input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, training the image segmentation model to be trained, and continuing to train the image segmentation model after training.
By the method, when the trained model is tested, the single-parameter magnetic resonance sample image can be used, so that the influence of each sub-convolution on the final result is adaptively adjusted under the condition that only the single-parameter mammary gland magnetic resonance image is input, the input can be processed most effectively, and accurate mammary gland lesion region segmentation is realized.
In the prior art, for the segmentation of a multi-parameter magnetic resonance image, a special network module needs to be designed to extract and fuse information provided by different imaging parameters, and the design of the special module is complex and the universality between different data sets is generally poor. In addition, the method also provides a well-registered multi-parameter magnetic resonance image during testing, and the testing cost is high. The method of the embodiment of the application can realize dynamic self-adaptive extraction of the multi-parameter magnetic resonance image information, does not need the multi-parameter image during testing, and ensures the accuracy of segmentation while reducing the testing cost.
In a second aspect of the embodiments of the present application, there is provided a multi-parameter breast magnetic resonance image segmentation apparatus based on a dynamic adaptive network, referring to fig. 4, including:
a training image obtaining module 401, configured to obtain a multi-parameter magnetic resonance sample image, where the multi-parameter magnetic resonance sample image includes a magnetic resonance sample image including a breast lesion region;
the model training module 402 is 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, where the image segmentation model to be trained is a dynamic adaptive network model;
a test image acquisition module 403, configured to acquire a single-parameter magnetic resonance sample image;
a model testing module 404, 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 the trained image segmentation model;
an image to be identified acquiring module 405, configured to acquire a magnetic resonance image to be identified;
and the to-be-identified image identification module 406 is configured to input the to-be-identified magnetic resonance image into the trained image segmentation model, so as to obtain information of the breast lesion area in the to-be-identified magnetic resonance image.
Optionally, the model training module 402 includes:
the identification result acquisition sub-module is used for inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
the training loss calculation submodule is used for calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and the model parameter adjusting submodule is used for adjusting parameters of the image segmentation model to be trained according to the current training loss, returning to input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
Optionally, the model testing module 404 includes:
the test loss calculation submodule is used for inputting 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;
and the model output submodule is used for obtaining a trained image segmentation model when the current test loss is less than a preset threshold value.
Optionally, the apparatus further comprises:
and the continuous training module is used for returning to input the multi-parameter magnetic resonance sample image into the image segmentation model to be trained when the current loss is greater than the preset threshold value, training the image segmentation model to be trained, and continuing to train the step of obtaining the trained image segmentation model.
Therefore, by the device provided by the embodiment of the application, the image segmentation model to be trained can be trained through the multi-parameter magnetic resonance sample image, and the trained model is tested through the single-parameter magnetic resonance sample image, so that the problem that the model can only be trained and tested through the multi-parameter image is solved, and the requirement on the sample image in the training process of the image segmentation model is reduced.
An embodiment of the present invention further provides an electronic device, as shown in fig. 5, which includes a processor 501, a communication interface 502, a memory 503 and a communication bus 504, where the processor 501, the communication interface 502 and the memory 503 complete mutual communication through the communication bus 504,
a memory 503 for storing a computer program;
the processor 501, when executing the program stored in the memory 503, implements the following steps:
acquiring a multi-parameter magnetic resonance sample image;
inputting a multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training the image segmentation model to be trained to obtain a trained image segmentation model;
and acquiring a single-parameter magnetic resonance sample image, inputting the single-parameter magnetic resonance sample image into the trained image segmentation model, and testing the trained image segmentation model to obtain the trained image segmentation model.
The communication bus mentioned in the electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the electronic equipment and other equipment.
The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components.
In yet another embodiment of the present invention, a computer-readable storage medium is further provided, in which a computer program is stored, and the computer program, when executed by a processor, implements any of the above dynamic adaptive network-based multi-parameter breast magnetic resonance image segmentation methods.
In yet another embodiment of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, causes the computer to perform any one of the above-mentioned embodiments of the dynamic adaptive network-based multi-parameter breast magnetic resonance image segmentation method.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may 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 loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. The term "comprising", without further limitation, means that the element so defined is not excluded from the group consisting of additional identical elements in the process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus, the electronic device, the storage medium and the computer program product embodiment, since they are substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (10)
1. A multi-parameter mammary gland magnetic resonance image segmentation method based on a dynamic self-adaptive network is characterized by comprising the following steps:
acquiring a multi-parameter magnetic resonance sample image, wherein the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region;
inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training 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 self-adaptive network model;
acquiring a single-parameter magnetic resonance sample image;
inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and obtaining the trained image segmentation model when the test is successful;
acquiring a magnetic resonance image to be identified;
and inputting the magnetic resonance image to be identified into the trained image segmentation model to obtain the information of the breast lesion area in the magnetic resonance image to be identified.
2. The method of claim 1, wherein the inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and training the image segmentation model to be trained to obtain a trained image segmentation model comprises:
inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and adjusting parameters of the image segmentation model to be trained according to the current training loss, returning to the image segmentation model to be trained, inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
3. The method of claim 1, wherein inputting the single-parameter magnetic resonance sample image into a trained image segmentation model tests the trained image segmentation model, and when the test is successful, obtaining the trained image segmentation model comprises:
inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to obtain the current test loss of the trained image segmentation model;
and when the current test loss is less than a preset threshold value, obtaining a trained image segmentation model.
4. The method of claim 3, wherein the inputting the single-parameter magnetic resonance sample image into the trained image segmentation model tests the trained image segmentation model, and when the testing is successful, after obtaining the trained image segmentation model, the method further comprises:
and when the current loss is larger than a preset threshold value, returning 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 continuing to train the step of obtaining the trained image segmentation model.
5. A multi-parameter mammary gland magnetic resonance image segmentation device based on a dynamic self-adaptive network is characterized by comprising:
the system comprises a training image acquisition module, a data acquisition module and a data processing module, wherein the training image acquisition module is used for acquiring a multi-parameter magnetic resonance sample image, and the multi-parameter magnetic resonance sample image comprises a magnetic resonance sample image containing a breast lesion region;
the model training module is used for 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 the trained image segmentation model, wherein the image segmentation model to be trained is a dynamic self-adaptive network model;
the test image acquisition module is used for acquiring a single-parameter magnetic resonance sample image;
the model testing module is used for inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to test the trained image segmentation model, and when the test is successful, the trained image segmentation model is obtained;
the to-be-identified image acquisition module is used for acquiring a to-be-identified magnetic resonance image;
and the to-be-identified image identification module is used for inputting the to-be-identified magnetic resonance image into the trained image segmentation model to obtain the information of the breast lesion area in the to-be-identified magnetic resonance image.
6. The apparatus of claim 5, wherein the model training module comprises:
the identification result acquisition submodule is used for inputting the multi-parameter magnetic resonance sample image into an image segmentation model to be trained, and identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained to obtain a current identification result;
the training loss calculation submodule is used for calculating the current training loss of the image segmentation model to be trained according to the recognition result;
and the model parameter adjusting submodule is used for adjusting parameters of the image segmentation model to be trained according to the current training loss, returning the multi-parameter magnetic resonance sample image to the image segmentation model to be trained, identifying the multi-parameter magnetic resonance sample image through dynamic convolution in the image segmentation model to be trained, and continuously executing the step of obtaining the current identification result until the preset iteration number is reached to obtain the trained image segmentation model.
7. The apparatus of claim 5, wherein the model test module comprises:
the test loss calculation submodule is used for inputting the single-parameter magnetic resonance sample image into a trained image segmentation model to obtain the current test loss of the trained image segmentation model;
and the model output submodule is used for obtaining a trained image segmentation model when the current test loss is less than a preset threshold value.
8. The apparatus of claim 7, further comprising:
and the continuous training module is used for returning to the step of inputting the multi-parameter magnetic resonance sample image into the image segmentation model to be trained when the current loss is greater than a preset threshold value, training the image segmentation model to be trained, and continuously training the step of obtaining the trained image segmentation model.
9. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any one of claims 1 to 5 when executing a program stored in the memory.
10. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111566049.XA CN114463345A (en) | 2021-12-20 | 2021-12-20 | Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network |
| PCT/CN2022/139816 WO2023116585A1 (en) | 2021-12-20 | 2022-12-17 | Multi-parameter breast magnetic resonance image segmentation method based on dynamic adaptive network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111566049.XA CN114463345A (en) | 2021-12-20 | 2021-12-20 | Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114463345A true CN114463345A (en) | 2022-05-10 |
Family
ID=81405429
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111566049.XA Pending CN114463345A (en) | 2021-12-20 | 2021-12-20 | Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN114463345A (en) |
| WO (1) | WO2023116585A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023116585A1 (en) * | 2021-12-20 | 2023-06-29 | 中国科学院深圳先进技术研究院 | Multi-parameter breast magnetic resonance image segmentation method based on dynamic adaptive network |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111539930B (en) * | 2020-04-21 | 2022-06-21 | 浙江德尚韵兴医疗科技有限公司 | Dynamic ultrasonic breast nodule real-time segmentation and identification method based on deep learning |
| CN112132808B (en) * | 2020-09-23 | 2023-11-07 | 沈阳建筑大学 | Mammary X-ray image lesion detection method and device based on normal model learning |
| CN113808143A (en) * | 2021-09-06 | 2021-12-17 | 沈阳东软智能医疗科技研究院有限公司 | Image segmentation method and device, readable storage medium and electronic equipment |
| CN114463345A (en) * | 2021-12-20 | 2022-05-10 | 中国科学院深圳先进技术研究院 | Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network |
-
2021
- 2021-12-20 CN CN202111566049.XA patent/CN114463345A/en active Pending
-
2022
- 2022-12-17 WO PCT/CN2022/139816 patent/WO2023116585A1/en unknown
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023116585A1 (en) * | 2021-12-20 | 2023-06-29 | 中国科学院深圳先进技术研究院 | Multi-parameter breast magnetic resonance image segmentation method based on dynamic adaptive network |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2023116585A1 (en) | 2023-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110222791B (en) | Sample labeling information auditing method and device | |
| CN110717881A (en) | Wafer defect identification method and device, storage medium and terminal equipment | |
| CN110570390B (en) | Image detection method and device | |
| CN111062426A (en) | Method, device, electronic equipment and medium for establishing training set | |
| CN113409284B (en) | Circuit board fault detection method, device, equipment and storage medium | |
| CN111915567A (en) | Image quality evaluation method, device, equipment and medium | |
| CN115294063A (en) | Method, device, system, electronic device and medium for detecting defect of electronic component | |
| CN112580668A (en) | Background fraud detection method and device and electronic equipment | |
| CN111339137A (en) | Data verification method and device | |
| CN110111311B (en) | Image quality evaluation method and device | |
| CN114463345A (en) | Multi-parameter mammary gland magnetic resonance image segmentation method based on dynamic self-adaptive network | |
| CN114998369A (en) | Image processing method, image processing device, storage medium and electronic equipment | |
| CN109064464B (en) | Method and device for detecting burrs of battery pole piece | |
| CN112435266A (en) | Image segmentation method, terminal equipment and computer readable storage medium | |
| CN112434717B (en) | Model training method and device | |
| CN114638304A (en) | Training method of image recognition model, image recognition method and device | |
| CN113158773B (en) | Training method and training device for living body detection model | |
| CN115458100A (en) | Knowledge graph-based follow-up method and device, electronic equipment and storage medium | |
| CN108764314B (en) | Structured data classification method and device, electronic equipment and storage medium | |
| CN114648646B (en) | Image classification method and device | |
| CN114882482B (en) | Seal anti-counterfeiting identification method and device | |
| CN112836819B (en) | Neural network model generation method and device | |
| CN110991460A (en) | Image recognition processing method, device, equipment and storage medium | |
| CN113297991A (en) | Behavior identification method, device and equipment | |
| CN116258873A (en) | Position information determining method, training method and device of object recognition model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |