Primary Transform Refinement for Cross-Component Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in refining transform coefficients, particularly in cross-component level reconstruction, which affects compression efficiency and quality.
Innovation Solution
Implementing a cross-component level reconstruction (CCLR) method that refines transform coefficients by adding offset values derived from co-located coefficients in other color components, using a target transform kernel for reverse transformation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If transform coefficients are refined using cross-component level reconstruction, then compression efficiency and quality are enhanced, but computational complexity increases
Solution Approach 1:
The patent introduces an intermediary refinement process that operates on transform coefficients using offset values derived from co-located coefficients in other color components. This intermediary step bridges the gap between basic transform coding and final reconstruction, enabling improved compression efficiency through cross-component information utilization while maintaining a modular architecture that manages computational complexity.
Solution Approach 2:
The patent modifies transform coefficients by adding offset values derived from cross-component information. This parameter change approach allows the system to enhance reconstruction quality by adjusting coefficient values based on correlations between different color components, thereby improving compression efficiency without requiring complete reprocessing of the entire coding pipeline.
2Manufacturing precision
If transform coefficients are refined by adding offset values from other color components, then reconstruction quality improves, but processing time increases
Solution Approach 1:
The patent performs preliminary refinement of transform coefficients by adding offset values derived from co-located coefficients in other color components before the final inverse transform and reconstruction steps. This preliminary action ensures that quality improvements are achieved early in the decoding process, allowing subsequent processing steps to operate on already-refined data and thereby reducing overall processing time.
Solution Approach 2:
The patent utilizes co-located transform coefficients from other color components as offset values to refine the current color component's coefficients. This copying approach leverages the temporal and spatial redundancy present in multi-color-component video data, improving reconstruction quality by transferring useful information from one color component to another without requiring additional source data or complex computations.
3Manufacturing precision
If cross-component level reconstruction is applied to all transform blocks, then overall video quality improves, but bitrate increases
Solution Approach 1:
The patent applies cross-component level reconstruction selectively rather than uniformly across all transform blocks. By identifying specific blocks where cross-component information provides the most benefit and applying the refinement process only to those blocks, the system achieves overall video quality improvement while minimizing the additional bitrate required, as not all blocks require the enhanced processing.
Data Source
AI summary
This disclosure relates generally to video coding and particularly to cross component level reconstruction. For example, a method is disclosed for processing video data which may include encoding a video block comprising a first transform block of a first color component and a second transform block of a second color component, wherein the first transform block and the second transform blocks are co-located blocks; encoding a first syntax element to signal whether a CCLR is applied to the first transform block, wherein, when CCLR is applied to the first transform block, a refinement process is required to be applied to the first transform block to obtain a refined first transform block on which an inverse transform is to be performed; and transmitting a bitstream comprising the video block and the first syntax element.