Low-dose medical image processing method, device and platform and storage medium
By performing generative adversarial network processing on low-dose medical image files, the problem of poor processing of traditional methods is solved, high-quality image processing is achieved, the accuracy of diagnosis is improved and the radiation dose is reduced.
Patent Information
- Application Number
- CN202510093939.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-21
- Publication Date
- 2025-05-23
AI Technical Summary
Traditional medical imaging methods have poor effect on low-dose medical imaging, resulting in low image quality and affecting the accuracy of diagnosis.
By obtaining the target low-dose medical image files, converting them to Mat-type files, and processing them based on the trained generative adversarial network, such as Super Resolution Generative Adversarial Network (SRGAN), to remove noise and improve image quality.
Effectively remove noise from low-dose medical images, retain key medical details, improve image quality to close to or reach high-dose imaging levels, improve diagnosis accuracy, and reduce radiation doses received by patients.
Smart Images
Figure CN120032819A_ABST
Abstract
Claims
1. A low-dose medical image processing method, characterized in that: include: Obtain target low-dose medical imaging files and record them as original files; Determining the file category of the target low-dose medical image file according to the file suffix of the target low-dose medical image file; According to the file category, converting the target low-dose medical image file into a Mat type file, converting the Mat type file into a QImage type file, and saving the original file, the Mat type file and the QImage type file respectively; The Mat type file is processed based on the trained generative adversarial network, the saved Mat type file is updated based on the processing result, and the saved QImage type file is updated based on the updated Mat type file to obtain the image processing result of the target low-dose medical image file.
2. The low-dose medical image processing method according to claim 1, characterized in that: The generative adversarial network is a super-resolution generative adversarial network or a cyclic generative adversarial network.
3. The low-dose medical image processing method according to claim 2, characterized in that: When the generative adversarial network is a super-resolution generative adversarial network, the objective function for training the super-resolution generative adversarial network is: The loss function of the generator in the super-resolution generative adversarial network is: The loss function of the discriminator in the super-resolution generative adversarial network is: Among them, G is the generator, D is the discriminator, V(G,D) is the objective function, is the expected value of the real data distribution, D(x) is the judgment output of the discriminator on the real image, is the expected value of the distribution of generated image data, G(x) is the output of the generator after processing the original image, G_loss is the loss function of the generator, N is the number of samples, x n is the nth sample, and D_loss is the loss function of the discriminator.
4. The low-dose medical image processing method according to claim 1, characterized in that: Also includes: The Mat type file is processed based on at least one of Gaussian filtering, mean filtering, grayscale transformation enhancement, edge sharpening and histogram equalization.
5. The low-dose medical image processing method according to claim 1, characterized in that: After obtaining the image processing result of the target low-dose medical image file, the method further includes: The image processing result is evaluated based on at least one of a mean square error, a peak signal-to-noise ratio, and a structural similarity index to obtain an evaluation result of the image processing result.
6. The low-dose medical image processing method according to claim 5, characterized in that: Also includes: Detecting whether the manual evaluation input field of the image processing result is valid; If the manual evaluation input column is valid, the manual evaluation result in the manual evaluation input column is saved as a txt file.
7. The low-dose medical image processing method according to claim 1, characterized in that: Also includes: In response to a request to switch to a learning and communication page, a knowledge map section and a communication and discussion section on low-dose medical image processing are displayed; In response to a request to switch to the knowledge map section, displaying a technology stack used to obtain an image processing result of the target low-dose medical image file and various technical courses on low-dose medical image processing; In response to a request to switch to the communication and discussion board, a community website is displayed, where the community website is used for question and answer and communication about low-dose medical imaging processing.
8. A low-dose medical image processing device, characterized in that: include: An acquisition module is used to acquire a target low-dose medical imaging file, which is recorded as an original file; A first processing module, configured to determine a file category of the target low-dose medical image file according to a file suffix of the target low-dose medical image file; A second processing module is used to convert the target low-dose medical image file into a Mat type file according to the file category, convert the Mat type file into a QImage type file, and save the original file, the Mat type file and the QImage type file respectively; The third processing module is used to process the Mat type file based on the trained generative adversarial network, update the saved Mat type file based on the processing result, and update the saved QImage type file based on the updated Mat type file to obtain the image processing result of the target low-dose medical image file.
9. A low-dose medical image processing platform, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method as claimed in any one of claims 1 to 7 are implemented.