Adaptive Block Filtering for Video Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Compression of digital video and images often results in loss of information, leading to noticeable distortions or 'blocking artifacts' along the borders of image blocks due to the truncation of higher frequency transform coefficients during encoding.
Innovation Solution
A block filtering system that identifies pairs of image data blocks separated by a boundary, determines a filter length based on edge contents, and applies filtering along the boundary using a multi-tap filter bank to reduce or remove blocking artifacts while preserving image edges and details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If image data is compressed using transform coefficients and quantization, then file size is reduced, but blocking artifacts appear along block boundaries
Solution Approach 1:
The patent applies different filtering strengths to different regions of the image based on local characteristics. The filter strength is adjusted according to the distance from block boundaries and the local edge content, allowing strong filtering in smooth regions while preserving edges near boundaries. This local adaptation resolves the contradiction by applying compression-friendly processing where needed while protecting important image features.
Solution Approach 2:
The filtering process is made dynamic by adaptively selecting filter lengths and strengths based on local image characteristics. The filter length varies from 1 to 4 pixels depending on the distance from the block boundary and the presence of edges. This dynamic approach allows the system to optimize between artifact reduction and detail preservation for each local region, resolving the contradiction between compression efficiency and visual quality.
2Object-generated harmful factors
If strong filtering is applied along block boundaries, then blocking artifacts are reduced, but image edges and details are blurred
Solution Approach 1:
The patent implements region-specific filtering by analyzing edge content at different distances from block boundaries. For pixels close to boundaries, the filter length is reduced or skipped entirely if edges are detected. For pixels farther from boundaries in smooth regions, full-strength filtering is applied. This local quality differentiation resolves the contradiction by protecting edges while removing artifacts in appropriate regions.
Solution Approach 2:
The filter length is dynamically adjusted based on the distance from the block boundary and local edge characteristics. The system transitions from strong filtering far from boundaries to weak or no filtering near boundaries with edges. This dynamic adaptation allows the system to resolve the contradiction between artifact reduction and edge preservation by optimizing filter parameters for each local context.
3Object-generated harmful factors
If filtering is applied to all pixels along block boundaries, then blocking artifacts are reduced, but computational complexity increases
Solution Approach 1:
The patent applies filtering selectively rather than uniformly to all boundary pixels. Filtering is applied with full strength only to pixels at certain distances from boundaries, while pixels closer to boundaries receive reduced or no filtering. This partial action approach reduces computational complexity compared to universal filtering while still effectively addressing blocking artifacts in the regions where they most occur.
Solution Approach 2:
The system performs local analysis of edge content to determine where filtering is needed and how strong it should be. By identifying regions with edges near block boundaries and excluding them from filtering, the system avoids unnecessary computation in those regions. This local quality assessment resolves the contradiction by applying computational resources only where they provide benefit.
Data Source
AI summary
A method includes identifying a pair of image data blocks separated by a boundary. The image data blocks include image information defining multiple pixels in at least one image. The method also includes identifying at least one filter length based on edge contents of at least some of the pixels in the at least one image. In addition, the method includes filtering at least some of the pixels in the pair of image data blocks along the boundary using the at least one identified filter length.