Multi-scale Contrast Enhancement Algorithm for CT Image Edge Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-scale image processing methods distort grey value transitions, leading to unnatural appearances and artifacts, especially in CT images with sharp transitions, due to excessive non-linearity in conversion functions.
Innovation Solution
A new multi-scale contrast enhancement algorithm that adjusts pixel-wise amplification functions based on the ratio of enhanced to unenhanced center differences, allowing for non-linear modification of translation difference images before reconstruction to preserve edge transitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If conversion functions are excessively non-linear to enhance contrast, then contrast enhancement is improved, but grey value transitions are distorted leading to unnatural appearance and artifacts
Solution Approach 1:
The patent segments the image processing into multiple scales using multi-scale decomposition, separating the image into detail images at different resolutions. This allows contrast enhancement to be applied selectively to different frequency components, preventing excessive non-linearity from distorting overall grey value transitions while still achieving local contrast enhancement where needed.
Solution Approach 2:
The patent applies different conversion functions to different detail images at different scales. By adjusting the non-linearity parameter individually for each scale's detail image, the method achieves local contrast enhancement in regions where it benefits the image while preserving grey value transition accuracy in regions where strong non-linearity would create artifacts.
2Illumination intensity
If strong non-linear conversion is applied to enhance contrast in CT images, then contrast enhancement is improved, but artifacts become more significant at sharp grey level transitions
Solution Approach 1:
The patent decomposes the CT image into multiple detail images at different scales, separating sharp transitions (which contain most of the artifact-prone information) from smoother regions. By applying controlled non-linear conversion at each scale rather than to the entire image, the method enhances contrast while limiting artifact generation at sharp grey level transitions.
Solution Approach 2:
The patent dynamically adjusts the non-linearity parameter of conversion functions based on the specific characteristics of each detail image at different scales. This dynamic adaptation allows the system to apply stronger non-linearity where it enhances contrast without creating artifacts, and weaker non-linearity where sharp transitions would be distorted, thereby reducing artifacts while maintaining contrast enhancement.
3Illumination intensity
If multi-scale decomposition is applied to enhance contrast, then contrast enhancement is improved, but computational complexity increases
Solution Approach 1:
The patent uses multi-scale decomposition to segment the image into a manageable number of detail images at different resolutions. This segmentation allows contrast enhancement to be performed on smaller, more manageable components rather than the entire high-resolution image, reducing overall computational complexity while maintaining enhancement effectiveness.
Solution Approach 2:
The patent applies contrast enhancement selectively to detail images at specific scales rather than processing all detail images at all scales. By identifying and enhancing only the scales that contribute most to perceptible contrast improvement, the method achieves effective contrast enhancement with reduced computational effort compared to processing the complete multi-scale representation.
Data Source
AI summary
A method of generating a multiscale contrast enhanced image is described wherein the shape of edge transitions is preserved. Detail images are subjected to a conversion, the conversion function of at least one scale being adjusted for each detail pixel value according to the ratio between the combination of the enhanced center differences and the combination of the unenhanced center differences.