Cross-Component Sample Offset Filtering for Video Reconstruction Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in effectively reducing reconstruction errors in video data compression, particularly in handling cross-component offsets between different color components, which can impact video quality and compression efficiency.
Innovation Solution
Implementing a cross-component offset filtering method that adjusts reconstruction values of a second color component using offset values derived from collocated and neighboring reconstructed samples of a first color component, utilizing a loop filter to enhance compression efficiency and reduce errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy compression is used to reduce bandwidth and storage requirements, then compression ratio is improved, but video quality and reconstruction accuracy deteriorate
Solution Approach 1:
The patent implements in-loop filtering where the filtered reconstructed samples are fed back into the prediction process. The offset calculation uses feedback from reconstructed luma samples to adjust chroma samples, creating a closed-loop system that continuously improves reconstruction accuracy while maintaining compression efficiency.
Solution Approach 2:
The patent dynamically changes parameters by calculating offsets based on gradients of luma samples and applying these offsets to chroma samples. The filter strength and offset values are adjusted based on local image characteristics, allowing adaptive reconstruction accuracy improvement without fixed compression loss.
2Manufacturing precision
If cross-component offset filtering is applied to improve reconstruction accuracy, then video quality is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent applies offset filtering locally by calculating gradients and offsets for specific sample positions rather than uniformly processing the entire image. The filter operates on individual chroma samples using local luma sample gradients, reducing overall computational complexity while maintaining accuracy where needed.
Solution Approach 2:
The patent segments the filtering process into distinct stages: gradient calculation from luma samples, offset computation, and application to chroma samples. This segmentation allows for optimized implementation where each stage can be independently processed and parallelized, reducing overall processing complexity.
3Manufacturing precision
If in-loop filtering is used to reduce reconstruction error, then video quality is improved, but bandwidth and storage requirements increase
Solution Approach 1:
The patent introduces offset values as intermediary data that mediates between the compressed chroma samples and the final reconstructed output. These offset values are calculated from luma samples and applied to chroma samples, improving reconstruction accuracy without requiring additional transmitted data beyond what is already in the bitstream.
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 syntax element for a CCSO mode. When the CCSO mode is enabled, a plurality of candidate luma sets are identified in a filter range that includes a first luma sample and neighboring luma samples. Each candidate luma set includes respective luma samples having positions symmetric with respect to a position of the first luma sample, and each luma sample located in the filter range is used in at least one of the candidate luma sets. A set of target luma samples is selected from the candidate luma sets. A loop filter is applied to combine the set of target luma samples and the first luma sample to generate the first sample offset of a first color sample collocated with the first luma sample.