Adaptive Illumination Compensation in Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding and decoding methods fail to effectively account for illumination variations between reference blocks and current blocks, particularly in bi-predicted scenarios, leading to suboptimal compression efficiency due to the use of a single illumination compensation flag for the entire coding unit.
Innovation Solution
The method involves determining and signaling illumination compensation parameters for each motion vector, allowing for separate illumination compensation flags and parameters for each reference picture list, thereby integrating illumination compensation into the motion field information, which is used for each prediction unit, enabling flexible and efficient illumination compensation during temporal prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single illumination compensation flag is used for the entire coding unit, then device complexity is reduced, but illumination compensation accuracy deteriorates
Solution Approach 1:
The patent segments the coding unit into multiple prediction units, each with its own motion vector and illumination compensation parameters. This allows independent illumination compensation for each prediction unit, improving accuracy while maintaining manageable complexity through structured organization.
Solution Approach 2:
The patent applies different illumination compensation parameters to different spatial regions (prediction units) within the coding unit. Each region can have its own slope and intercept parameters, allowing local adaptation to illumination variations while preserving overall system efficiency.
2Measurement precision
If separate illumination compensation parameters are determined for each motion vector, then illumination compensation accuracy is improved, but encoding and decoding time increases
Solution Approach 1:
The patent determines illumination compensation parameters for each motion vector during the motion estimation phase, before final encoding. This preliminary determination allows the parameters to be cached and reused, reducing the computational burden during actual encoding and decoding operations.
Solution Approach 2:
The patent introduces illumination compensation parameters (slope and intercept) that modify the prediction block values. These parameters are derived from reference block statistics and applied efficiently through linear transformation, balancing accuracy improvement with computational efficiency.
Data Source
AI summary
Different implementations are described for determining one or more illumination compensation parameters for a current block being encoded by a video encoder or decoded by a video decoder. A plurality of motion vectors for a current block being encoded in a picture are determined. One or more illumination compensation parameters for each of the plurality of motion vectors are determined and encoded or decoded. The current block is then encoded or decoded using the plurality of motion vectors and the one or more illumination compensation parameters for each of the plurality of motion vectors. In one embodiment, a flag is used to signal the use or not of the illumination compensation. In another embodiment, the illumination compensation flag is not encoded or decoded if illumination compensation is not used.