Unified Merge Candidate List for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies require separate merge modes and candidate lists for translational and affine motion models, leading to increased complexity and reduced coding efficiency due to the need for additional flags and parsing operations.
Innovation Solution
A unified merge mode candidate list is proposed, which combines translational and affine motion vector candidates, eliminating the need for separate flags and lists, and allowing for more efficient decoding and encoding processes by using a single merge index to select motion vectors from a unified candidate list.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate merge modes and candidate lists are used for translational and affine motion models, then motion prediction accuracy is improved, but device complexity and coding efficiency deteriorate due to additional flags and parsing operations
Solution Approach 1:
The patent combines separate translational and affine motion model candidate lists into a single unified merge candidate list. The decoder receives a single merge index that directly points to candidates in this unified list, eliminating the need for separate flags and lists while maintaining support for both motion models. This reduces parsing operations and device complexity.
Solution Approach 2:
The unified merge candidate list serves multiple functions by containing both translational and affine motion vector candidates. A single merge index can select from either type of candidate depending on its position in the unified list, making the candidate list structure universal for both motion models rather than requiring separate specialized lists.
2Measurement precision
If separate merge modes and candidate lists are used for translational and affine motion models, then motion prediction accuracy is improved, but coding efficiency deteriorates due to additional flags and parsing operations
Solution Approach 1:
The patent combines separate translational and affine motion model candidate lists into a single unified merge candidate list. The decoder receives a single merge index that directly points to candidates in this unified list, eliminating the need for separate flags and lists while maintaining support for both motion models. This reduces parsing operations and device complexity.
3Measurement precision
If separate merge modes and candidate lists are used for translational and affine motion models, then motion prediction accuracy is improved, but ease of operation deteriorates due to additional flags and parsing operations
Solution Approach 1:
The patent combines separate translational and affine motion model candidate lists into a single unified merge candidate list. The decoder receives a single merge index that directly points to candidates in this unified list, eliminating the need for separate flags and lists while maintaining support for both motion models. This reduces parsing operations and device complexity.
Data Source
AI summary
A video decoder including one or more processors configured to receive one or more bits, in a bitstream, that indicate the encoded current block of video data was encoded based on a unified candidate list that includes motion vector candidates based on one or more translational motion vectors, and motion vector candidates based on one or more affine motion vectors. A merge index represented in the bitstream may indicate which candidate in the unified candidate list is associated with the motionvector of the encoded current block of video data. Based on the merge index, the one or more processors are configured to select one or more motion vectors of a candidate from the unified candidate list, based on the merge index, where the candidate has one or more of the motion vectors corresponding to the translational motion vectors or affine motion vectors within the unified candidate list.