UMVE-Based Pruning for Video Motion Vector Candidate Lists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques face challenges in efficiently constructing motion vector candidate lists, leading to suboptimal motion vector prediction and resulting in reduced video quality and increased bitrate.
Innovation Solution
The implementation of an ultimate motion vector expression (UMVE)-based pruning method, which expands potential motion vector candidates by modifying existing candidates with distances and directions, allowing for more accurate motion vector prediction and improved list construction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional motion vector candidate list construction is used, then the coding process is simpler, but the motion vector prediction accuracy is reduced leading to higher bitrate
Solution Approach 1:
The patent applies preliminary action by pre-expanding motion vector candidates using UMVE (Ultimate Motion Vector Expression) before the actual candidate list construction. By modifying existing motion vector candidates with distances and directions in advance, the system creates a richer pool of potential candidates that can be directly used during encoding, improving prediction accuracy without adding complex real-time processing during the candidate selection phase.
Solution Approach 2:
The patent uses copying by creating multiple UMVE candidate versions from each base motion vector candidate. Each original motion vector candidate is copied and modified with different distance and direction parameters to generate additional candidate variations. This copying approach increases the diversity of candidates available for prediction without requiring fundamentally new candidate generation mechanisms.
2Adaptability or versatility
If UMVE expansion is applied to all motion vector candidates, then more unique candidates are available, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies partial action by selectively expanding only certain motion vector candidates using UMVE rather than all candidates. The system evaluates base candidates and applies UMVE expansion to those that meet specific criteria or show potential for improvement, rather than uniformly processing every candidate. This selective approach maintains a good balance between generating diverse candidates and maintaining coding throughput.
Solution Approach 2:
The patent applies local quality by applying different levels of UMVE expansion to different candidates based on their characteristics. Some candidates receive full UMVE expansion with multiple distance and direction modifications, while others receive minimal or no expansion. This differentiated approach ensures that computational resources are focused on candidates where UMVE expansion provides the most value, improving overall candidate diversity without uniformly increasing processing complexity.
3Reliability
If redundant motion vector candidates are included in the candidate list, then the list is more complete, but the processing efficiency and coding throughput are reduced
Solution Approach 1:
The patent applies the taking out principle by extracting and removing redundant motion vector candidates from the candidate list after UMVE expansion. The system identifies candidates that are duplicates or highly similar to existing candidates and removes them from the final candidate list. This extraction process maintains the completeness and diversity of unique candidates while eliminating redundant entries that would otherwise reduce coding throughput and increase processing overhead.
Data Source
AI summary
A video encoder and video decoder are configured to perform an ultimate motion vector expression (UMVE)-based pruning method which is used to prune motion vectors in a motion vector candidate list. The video encoder and video decoder may add one or more motion vector candidates to a candidate list for motion vector prediction for a current block of the video data, determine whether to add a next motion vector candidate to the candidate list based on a UMVE candidate of a respective candidate of the one or more candidates, and encode/decode the current block of the video data using the candidate list.