Edge Enhancement Filter for AV1 Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AOMedia Video 1 (AV1) experiences information loss during video compression, particularly in edge detection, which affects coding efficiency and visual quality due to lossy coding modules and the limitations of current high-pass filters like CCALF, which can only be applied to chroma components without signaling filter coefficients.
Innovation Solution
The method involves detecting edges within video data, calculating gradient values for edge directions, and using these gradients to decode and encode video data, allowing filtering to be applied across all color components without needing to signal filter coefficients, thereby enhancing edge detection and preservation across all color components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CCALF filter is used for chroma components, then edge detection is improved, but filter coefficients need to be signaled increasing bitstream overhead
Solution Approach 1:
The system uses the video data itself to determine filtering parameters. Gradient calculations are performed on the decoded video samples to automatically identify edge directions and strengths, eliminating the need for external signaling of filter coefficients. The filter adapts to local content characteristics without requiring additional bitstream information.
Solution Approach 2:
The filtering parameters (coefficients, edge directions, strengths) are dynamically changed based on local gradient calculations rather than being fixed or pre-specified. The system computes gradients in multiple directions (horizontal, vertical, diagonal) and adjusts filter application accordingly, allowing parameters to adapt to local image characteristics without signaling overhead.
2Manufacturing precision
If filtering is applied to all color components, then visual quality is improved, but device complexity increases
Solution Approach 1:
The filtering process is segmented by color component (luma and chroma) and by directional gradients. Each component can be processed independently with appropriate gradient calculations, allowing selective application of filtering where needed without uniformly processing all data through complex filters.
Solution Approach 2:
The same gradient-based filtering mechanism is applied universally across all color components (Y, Cb, Cr or R, G, B). The filter serves multiple functions: edge detection, artifact reduction, and quality enhancement, all through a single unified process that works across different color spaces and components.
3Measurement precision
If gradient calculations are performed for all directions, then edge detection accuracy is improved, but processing time increases
Solution Approach 1:
The system calculates gradients in multiple directions (horizontal, vertical, and diagonal) but applies filtering only where gradients exceed certain thresholds or where edges are actually detected. Not all gradient calculations lead to filter application, avoiding unnecessary processing while maintaining accuracy where needed.
Solution Approach 2:
Gradient calculations and filtering are applied locally based on detected edge characteristics rather than uniformly across the entire frame. The filter strength, direction, and application are customized for each local region based on its specific gradient profile, improving accuracy while reducing overall processing time by focusing computation on relevant areas.
Data Source
AI summary
A method, computer program, and computer system is provided for coding video data. Video data is received, and an edge present within a sample of the received video data is detected. A gradient value corresponding to a direction associated with the detected edge is calculated. The video data is decoded based on the calculated gradient.