Decoder-Side Motion Vector Refinement for Selective Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in efficiently encoding and decoding video data due to limitations in motion vector prediction and refinement, particularly in advanced video coding formats like HEVC and VVC, leading to suboptimal bandwidth usage and decoding quality.
Innovation Solution
Implementing decoder-side motion vector derivation (DMVD) schemes that apply unequal weighting factors and refine motion information using weight parameters to improve the conversion between video blocks and bitstream representations, enabling or disabling these schemes based on picture order count, bitstream representations, and block dimensions, and applying symmetric motion vector difference modes to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If decoder-side motion vector derivation (DMVD) schemes are applied to all video blocks, then motion information refinement is improved, but decoding complexity and processing time increase
Solution Approach 1:
The patent applies DMVD selectively to specific video blocks based on local characteristics such as block size, motion complexity, and prediction mode. Rather than uniformly processing all blocks, the decoder identifies and applies refinement only where beneficial, reducing overall complexity while maintaining precision where needed.
Solution Approach 2:
The patent implements partial DMVD by applying motion vector derivation to only a subset of video blocks that meet specific criteria (e.g., large blocks, high motion areas). This partial application reduces processing load while still achieving sufficient refinement for quality-critical regions.
2Measurement precision
If unequal weighting factors are applied to prediction blocks, then prediction accuracy is improved, but coding complexity increases
Solution Approach 1:
The patent modifies the weighting parameters used in bi-prediction by introducing unequal weights based on local image characteristics. The weighting factors are adjusted according to motion vector differences, block positions, and reference picture qualities, allowing adaptive prediction accuracy without requiring complex重新calculation for every block.
3Stability of the object's composition
If symmetric motion vector difference (SMVD) mode is enabled for all blocks, then motion consistency is improved, but processing overhead increases
Solution Approach 1:
The patent segments video blocks into different categories based on their motion characteristics and applies SMVD selectively. By dividing the processing into segments (blocks requiring SMVD vs. blocks using standard modes), the system maintains motion consistency where needed while avoiding unnecessary processing overhead in other regions.
4Measurement precision
If DMVD is applied at sub-block level, then motion refinement precision is improved, but computational load increases significantly
Solution Approach 1:
The patent applies sub-block level DMVD only to specific sub-blocks within larger blocks that exhibit high motion variability or prediction errors. By performing partial sub-block processing rather than uniform full-block processing, the system achieves high precision where needed while significantly reducing overall computational load.
Data Source
AI summary
One example method of video processing includes implementing, by a processor, a decoder-side motion vector derivation (DMVD) scheme for motion vector refinement during a conversion between a current video block and a bitstream representation of the current video block by deriving parameters based on a deriving rule. The conversion may include compressing the current video block into the bitstream representation or uncompressing the bitstream representation into pixel values of the current video block.