Adaptive CCSO Filter Unit Sizing for Video Reconstruction Error Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently reducing reconstruction errors in video data while maintaining flexibility and adaptability in filter unit sizes for cross-component offset filtering.
Innovation Solution
Implementing a cross-component sample offset (CCSO) filter that uses co-located reconstructed samples and neighboring samples from a first color component to derive a sample offset value for a second color component, with adaptive filter unit sizes and control parameters for enhanced flexibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed filter unit size of 256×256 luma samples is used for CCSO filtering, then the device complexity is reduced and processing is simplified, but the adaptability to different video content and regions is limited
Solution Approach 1:
The picture is divided into multiple filter units of different sizes (e.g., 256×256, 128×128, 64×64 luma samples) instead of using a single fixed size. This segmentation allows the filtering process to adapt to different regional characteristics in the video content, improving adaptability while maintaining manageable complexity through a finite set of predefined size options.
Solution Approach 2:
The filter unit size is made dynamic and adjustable rather than fixed. The decoder can select different filter unit sizes based on video content characteristics, allowing the system to adapt its behavior to different scenarios. This dynamic adjustment resolves the contradiction by enabling adaptability without requiring completely complex custom configurations.
2Manufacturing precision
If adaptive filter unit sizes are implemented for CCSO filtering, then the reconstruction error reduction is improved and image quality is enhanced, but the processing complexity and computational load increase
Solution Approach 1:
Different filter unit sizes are applied to different regions of the picture based on local content characteristics. Important regions with detailed content use smaller filter units for more precise reconstruction, while less critical regions use larger filter units. This local quality approach improves overall reconstruction accuracy without uniformly increasing complexity across the entire processing system.
Solution Approach 2:
The filter unit size parameter is made variable and can be changed based on content analysis. By adjusting this single parameter (filter unit size) rather than completely redesigning the filtering architecture, the system achieves improved reconstruction accuracy with relatively controlled increases in processing complexity.
3Adaptability or versatility
If multiple filter unit sizes are supported for different color components, then the versatility and control flexibility are improved, but the signaling overhead and processing complexity increase
Solution Approach 1:
The same set of filter unit size options and control mechanisms are used across different color components (luma and chroma). This universal approach allows the system to handle multiple color components with a single, unified control structure, improving versatility while avoiding the need for separate complex signaling systems for each component.
Data Source
AI summary
Various implementations described herein include methods and systems for coding video. In one aspect, a video bitstream includes a current image frame and a first filter control parameter for a loop filter to process a first filtering block of the current image frame. An electronic device receives the video bitstream and determines a filter unit size for processing the current image frame by the loop filter. The first filtering block has the filter unit size. The first filtering block is identified in the current image frame based on the filter unit size. When the first filter control parameter is enabled, the loop filter is applied to process one or more samples of the first filtering block. The current image frame includes the first filtering block.