Video Block Prediction Using Multi-Candidate LIC Inheritance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding methods face challenges in efficiently handling motion information inheritance from multiple motion candidates, particularly when local illumination compensation (LIC) indicators are ambiguous due to differing values among candidates.
Innovation Solution
A method and apparatus for video encoding and decoding that determine a prediction for a block by inheriting motion information from at least two motion candidates, and adjust an indicator for enabling or disabling correction based on the values of corresponding indicators from these candidates, ensuring consistent and efficient use of LIC or other features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion information is inherited from multiple motion candidates, then compression efficiency is improved, but ambiguity in LIC indicators arises due to differing values among candidates
Solution Approach 1:
Instead of selecting one motion candidate's LIC indicator value, the patent inverts the approach by combining information from multiple candidates through a derivation process. The LIC indicator for the current block is derived by checking whether LIC is enabled for each motion candidate and applying a logical combination (e.g., logical AND or OR operations) to determine the final LIC indicator value, thereby resolving the ambiguity while maintaining compression efficiency.
Solution Approach 2:
The patent introduces an intermediary derivation process that acts as a mediator between multiple motion candidates with different LIC indicator values. This intermediary process systematically combines the LIC indicator information from multiple candidates through defined rules and operations, producing a consistent and unambiguous LIC indicator for the current block without requiring additional signaling.
2Measurement precision
If LIC correction is applied to improve prediction accuracy, then decoding quality is improved, but bitrate increases due to additional correction parameters
Solution Approach 1:
The patent enables the LIC correction mechanism to be self-service by deriving the LIC indicator and correction parameters automatically from the motion candidates' existing information. The system uses the LIC indicator values already present in the motion candidates to determine whether LIC correction should be applied and to derive the correction parameters, eliminating the need for separate signaling and reducing bitrate overhead while maintaining prediction accuracy.
Data Source
AI summary
A method for coding or decoding a block is provided, as well as a corresponding apparatus. A prediction for the block is determined, wherein motion information that is used for determining the prediction is inherited from motion information of at least two motion candidates, and the block is coded or decoded based on the determined prediction, wherein determining the prediction for the block comprises determining a value for an indicator enabling or disabling correction of the prediction, based on values of corresponding indicators of the at least two motion candidates.