Video Coding Tree Splitting Using Motion Field Homogeneity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current advanced block partitioning methods in video coding, such as multi-type tree structures, do not consider motion field information during tree splits, leading to inefficiencies as block partitioning often correlates with motion fields and results in increased bit rates.
Innovation Solution
A method that partitions video sequences into coding tree units, determining the motion field of each block and deciding whether to further partition based on its homogeneity or heterogeneity, using motion field information to optimize tree splits and reduce signaling costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If advanced block partitioning methods (such as multi-type tree structures) are used to increase partitioning flexibility, then block partitioning adaptability is improved, but bit rate increases due to not considering motion field information
Solution Approach 1:
The patent applies local quality by making the partitioning decision dependent on the local motion field characteristics of each block. Instead of using a uniform partitioning strategy, the method evaluates motion homogeneity specifically for each block and adapts the partitioning accordingly - using finer partitions only where motion varies significantly, while maintaining coarser partitions in regions with uniform motion, thus reducing overall bit rate while preserving necessary adaptability
Solution Approach 2:
The patent changes the partitioning parameters (tree depth, partition type) based on the motion field parameter (motion homogeneity). By dynamically adjusting partitioning parameters according to the measured motion characteristics, the system achieves adaptability without unnecessarily increasing bit rate, as partitioning is only refined when motion field analysis indicates it is beneficial
2Device complexity
If block partitioning is performed without considering motion field information, then encoding complexity is reduced, but coding efficiency deteriorates due to increased bit rate
Solution Approach 1:
The patent performs preliminary action by evaluating motion field homogeneity before making partitioning decisions. This preliminary motion analysis allows the encoder to anticipate which blocks will benefit from further partitioning, avoiding unnecessary partitioning operations and their associated signaling overhead, thus improving coding efficiency without substantially increasing encoding complexity
Solution Approach 2:
The motion field information serves the dual purpose of both analyzing block characteristics and guiding partitioning decisions. The same motion data used for motion compensation also informs the partitioning strategy, eliminating the need for separate complexity-intensive analysis and allowing the system to improve coding efficiency using already-computed information
3Loss of information
If motion field information is considered during tree splits, then bit rate is reduced through optimized partitioning, but encoding complexity increases due to additional motion analysis
Solution Approach 1:
The patent applies universality by making the motion field analysis serve multiple functions simultaneously - it is used for both motion compensation (a required operation in inter-prediction) and for guiding partitioning decisions. This multi-functionality allows the system to reduce bit rate through motion-aware partitioning without incurring additional complexity, as the motion information is already being computed for other purposes
Data Source
AI summary
A method and apparatus for encoding or decoding a video sequence includes partitioning the video sequence into coding tree units, each coding tree unit including at least one coding tree block and each coding tree block including at least one coding block, determining the motion field of the at least one coding block, determining whether the motion field of the at least one coding block is homogenous or heterogeneous, and determining whether to signal a further partition of the at least one coding block based on the determination of whether the motion field of the at least one coding block is homogenous or heterogeneous.