Image Sharpening via Bilateral and Gaussian Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image sharpening methods using high-pass filters like Laplacian often result in a halo effect, leading to poor sharpening effects due to overshoot and undershoot at bright-dark boundaries, which manifest as 'white margin' and 'black margin' at black-white edges.
Innovation Solution
The method involves performing bilateral filtering and difference of Gaussians filtering on original image information to obtain distinct image-layer information, which is then fused and superimposed to reduce the halo effect, improving the image sharpening effect by combining the characteristics of both filtering processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a high-pass filter (Laplacian) is used to extract high-frequency information and superimpose it on the original image, then the edge and outline enhancement effect is improved, but a halo effect occurs at bright-dark boundaries
Solution Approach 1:
The patent segments the image processing into multiple filtering stages: first applying bilateral filtering to obtain edge information, then applying difference of Gaussians filtering to obtain high-frequency information, and finally combining these segmented results. This segmentation allows each filter to perform its specialized function without generating harmful artifacts.
Solution Approach 2:
The patent applies different filtering operations to different aspects of the image locally: bilateral filtering is applied to preserve edge information while difference of Gaussians is applied to extract high-frequency details. By combining these locally processed results with appropriate weighting, the method enhances edges without creating halo effects at boundaries.
2Manufacturing precision
If bilateral filtering is applied to preserve edges, then edge clarity is improved, but noise may be amplified
Solution Approach 1:
The patent merges the results of bilateral filtering (which preserves edges) with the results of difference of Gaussians filtering (which extracts high-frequency information). By combining these two filtered results through addition or weighted combination, the method achieves both edge clarity and noise suppression, as each filter compensates for the weaknesses of the other.
3Manufacturing precision
If difference of Gaussians filtering is used to extract high-frequency information, then the sharpening effect is improved, but halo artifacts appear at high-contrast edges
Solution Approach 1:
The patent creates a composite processing result by combining the output of bilateral filtering with the output of difference of Gaussians filtering. This composite approach leverages the strengths of both filtering methods: bilateral filtering provides clean edge information without halo artifacts, while difference of Gaussians provides strong high-frequency enhancement. The combination achieves superior sharpening without the halo artifacts that would result from using difference of Gaussians alone.
Data Source
AI summary
Embodiments of the present invention provide an image sharpening method and device. The method includes performing bilateral filtering processing and difference of Gaussians filtering processing on original image information to obtain first image-layer information and second image-layer information respectively. The first image-layer information is subtracted from the original image information to obtain third image-layer information. Fusion and superimposition processing is performed on the second image-layer information and the third image-layer information to obtain fourth image-layer information. The original image information and the fourth image information are added to obtain processed image information.