Spatial Local Illumination Compensation for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding methods struggle to effectively compensate for illumination discrepancies between different regions or blocks within the same picture, particularly in scenarios with gradual spatial illumination variations, which affects coding efficiency.
Innovation Solution
The method involves determining parameters for local illumination compensation based on spatially neighboring reconstructed samples and corresponding samples of neighboring blocks within the same picture, using a linear model with scaling and offset factors to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional video coding schemes are used without illumination compensation, then the device complexity remains low, but the coding efficiency deteriorates due to inability to compensate illumination discrepancy between different regions
Solution Approach 1:
The patent applies local illumination compensation by dividing the video picture into multiple blocks and independently determining illumination parameters (offset and scaling factors) for each block based on its specific neighboring blocks. This allows different regions with varying illumination characteristics to be compensated locally, improving coding efficiency without requiring global complexity
Solution Approach 2:
The patent segments the video picture into multiple blocks and processes each block separately with its own illumination compensation parameters. By segmenting the processing into block-level operations using neighboring block samples, the solution achieves localized illumination correction that improves overall coding efficiency while maintaining manageable computational complexity through parallelizable block processing
2Measurement precision
If spatial local illumination compensation is applied, then the prediction accuracy improves, but the computational complexity increases due to additional parameter determination and processing
Solution Approach 1:
The patent performs preliminary action by determining illumination compensation parameters (offset and scaling factors) for each block before the actual prediction process. By pre-calculating these parameters using neighboring block samples and applying them to correct illumination discrepancies, the prediction accuracy is improved while the computational overhead is confined to a one-time parameter determination step
Solution Approach 2:
The patent introduces illumination compensation parameters (offset and scaling factors) that modify the prediction process by adjusting for illumination differences. These parameter changes are derived from neighboring block samples and applied to enhance prediction accuracy, adding computational steps but enabling precise compensation for spatial illumination variations that traditional methods cannot handle
Data Source
AI summary
At least a method and an apparatus are presented for efficiently encoding or decoding video. For example, parameters for a local illumination compensation LIC of a current block being encoded/decoded in a picture are determined based on spatially neighboring reconstructed samples and corresponding spatially neighboring reconstructed samples of at least one spatial reference block, wherein the at least one spatial reference block is a spatially neighboring block of the current block in the picture. For example, a flag enables/disables the spatial LIC for the current block. For example, the spatial LIC is applied to any of an Inter/Intra/IBC prediction. For example, multiple spatial reference blocks are used in determining the spatial LIC parameters. For example, spatially neighboring reconstructed samples of multiple lines are used in determining the spatial/temporal LIC parameters.