Multiple-Parameter Local Illumination Compensation for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies struggle to efficiently handle sequences with illumination changes, leading to inefficiencies in bandwidth usage and computational complexity.
Innovation Solution
Implement multiple-parameter local illumination compensation (MPLIC) to enhance video data processing, including methods such as extended merge prediction, high precision motion compensation, and affine motion prediction, to improve inter prediction in video coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing video coding technologies are used to handle sequences with illumination changes, then bandwidth usage and computational complexity increase, but video coding efficiency remains insufficient
Solution Approach 1:
The patent applies parameter changes by introducing multiple illumination compensation parameters (scale factors and offset values) to adaptively model illumination variations. These parameters are derived from template blocks and reference blocks, allowing the system to dynamically adjust to different illumination conditions without requiring additional bandwidth for full block data transmission.
Solution Approach 2:
The patent segments the current block into multiple sub-blocks, each with its own illumination compensation parameters. This segmentation allows localized illumination adjustment, improving coding efficiency for blocks with non-uniform illumination changes while maintaining lower overall bandwidth requirements compared to transmitting full high-resolution data.
2Measurement precision
If existing video coding technologies are used to handle sequences with illumination changes, then computational complexity increases, but prediction accuracy remains insufficient
Solution Approach 1:
The patent applies local quality by deriving separate illumination compensation parameters for different sub-blocks within a block. Each sub-block receives customized scale and offset values based on its local illumination characteristics, improving prediction accuracy for regions with varying illumination while keeping computational complexity manageable through localized processing.
Solution Approach 2:
The patent performs preliminary action by pre-calculating illumination compensation parameters from template blocks before the main decoding process. These pre-derived parameters are then applied during inter-prediction, reducing real-time computational complexity while maintaining high prediction accuracy for illumination-changed sequences.
Data Source
AI summary
A mechanism for processing video data is disclosed. The mechanism includes determining to apply a multiple-parameter local illumination compensation (MPLIC) to a visual media data. MPLIC can be used as an additional local illumination compensation (LIC) mode or instead of LIC. A conversion is performed between the visual media data and a bitstream based on the MPLIC.