Geometric Partitioning Weight Derivation for Video Motion Blending
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently handling geometric partitioning modes, particularly in deriving motion vectors and blending weights for improved compression efficiency.
Innovation Solution
The method involves determining a weight index at a sample position within a coding block partitioned into geometric parts, using this index to derive a blending weight for merging prediction signals and storing motion vectors based on a motion index, without relying on lookup tables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If lookup tables are used for deriving motion vectors and blending weights in geometric partitioning mode, then the accuracy of motion vector derivation is improved, but the device complexity and memory requirements increase
Solution Approach 1:
The patent extracts and eliminates the need for large lookup tables by deriving motion vectors and blending weights through direct mathematical calculations using only basic arithmetic operations and a small set of pre-computed constants, thereby reducing memory requirements while maintaining accuracy
Solution Approach 2:
The patent changes the approach from storing pre-computed values in lookup tables to calculating values on-demand using mathematical formulas, transforming the data representation from static tables to dynamic calculations with reduced memory footprint
2Productivity
If geometric partitioning mode is implemented with comprehensive motion vector derivation, then compression efficiency is improved, but computational complexity increases
Solution Approach 1:
The patent segments the coding block into geometric partitions (triangular or wedge-shaped regions) and processes each partition independently with simplified motion vector derivation, reducing the overall computational complexity while maintaining high compression efficiency through the geometric partitioning strategy
Solution Approach 2:
The patent applies partial motion vector derivation by computing motion vectors only for the geometric partitions that actually contain motion information, rather than uniformly processing the entire coding block, thereby reducing unnecessary computations while preserving compression efficiency
3Measurement precision
If explicit motion parameter transmission is used for each coding unit, then prediction accuracy is improved, but the bitrate increases
Solution Approach 1:
The patent enables the coding system to self-derive motion vectors and blending weights using locally available information and mathematical calculations, eliminating the need to explicitly transmit all motion parameters and reducing the bitrate while maintaining prediction accuracy
Solution Approach 2:
The patent introduces mathematical calculations and derived parameters as intermediaries between the raw video data and the prediction process, allowing accurate prediction to be achieved through computation rather than explicit parameter transmission, thereby reducing bitrate requirements
Data Source
AI summary
A method of video decoding including receiving a syntax element from a bitstream of a coded video associated with a current coding block that is partitioned into a first part and a second part along a partition edge for prediction in a geometric partitioning mode, and determining a weight index at a sample position of the current coding block. A stored motion vector type for a first motion storage unit in a motion field of the current coding block is determined based on a motion index of the first motion storage unit. The weight index at the sample position of the current coding block is used as the motion index of the first motion storage unit. The motion field of the current coding block can be stored based on the stored motion vector type determined for the first motion storage unit in the motion field of the current coding block.