CT image enhancement method and device based on global consistency optimization
The CT image enhancement method with global consistency optimization, based on Gaussian fitting and pixel label adjustment, solves the problem of noise suppression and detail preservation in low-dose CT imaging, and achieves high-quality imaging under low radiation, which is suitable for multiple application scenarios.
CN122089577AActive Publication Date: 2026-05-26SUN YAT SEN UNIV
8 Cites 0 Cited by
Patent Information
- Application Number
- CN202610220404.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-24
- Publication Date
- 2026-05-26
- Estimated Expiration
- 2046-02-24
AI Technical Summary
Technical Problem
Existing CT imaging techniques struggle to balance noise suppression and detail enhancement while reducing radiation dose, and their generalization ability is limited due to reliance on training data or equipment modifications.
Method used
By employing a global consistency optimization method, based on Gaussian fitting and pixel label adjustment, a global optimization objective function is constructed to adaptively reduce pixel intensity bias and generate enhanced CT images.
Benefits of technology
Significantly improves signal-to-noise ratio and structural clarity at low radiation doses, suppresses noise and preserves detail, suitable for industrial non-destructive testing, medical diagnostics and high-resolution microscopy, reducing radiation risks and costs.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses a CT image enhancement method and device based on global consistency optimization, and the method comprises the steps: carrying out the Gaussian fitting based on the pixel intensity distribution of a low-dose CT image, and obtaining a fitting result containing a plurality of types of Gaussian distribution components; calculating the category probability that each pixel intensity belongs to each Gaussian distribution component pixel by pixel based on the fitting result, and distributing a pixel label for each pixel; constructing a global optimization objective function containing the consistency constraint of adjacent pixel tags, adjusting the pixel tags through iterative optimization to maximize the objective function, and obtaining optimized pixel tags; and according to the mean value of the Gaussian distribution component corresponding to each optimized pixel label, carrying out adaptive deviation reduction processing of reducing the intensity value of each pixel to the mean value, and generating an enhanced CT image. According to the invention, noise suppression and detail maintenance can be realized at the same time under the condition of low radiation dose, and the method is suitable for industrial nondestructive testing, medical diagnosis, high-resolution microscopic imaging and other scenes.
Need to check novelty before this filing date? Find Prior Art