Implicit Masked Blending With MV Refinement at Reference Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods suffer from inaccurate prediction due to reference blocks extending beyond boundaries, leading to suboptimal padding and reduced video quality.
Innovation Solution
Implement an implicit masked blending mode that uses boundary-aware compound prediction with motion vector refinement, weighting real pixel data more heavily and refining motion vectors to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If reference blocks are used that extend beyond boundaries, then the prediction coverage is improved, but the prediction accuracy deteriorates due to suboptimal padding
Solution Approach 1:
The patent applies different processing strategies to different regions of the reference block based on their boundary status. In-bounds regions use standard prediction while out-of-bounds regions use gradient extrapolation, giving each region the quality treatment it needs to maintain overall accuracy
Solution Approach 2:
The patent pre-calculates gradients from available reference pixels before performing the actual prediction. This preliminary computation of gradient values allows for accurate extrapolation into out-of-bounds regions without degrading prediction quality
2Ease of manufacture
If padding is applied to out-of-boundary reference blocks, then the prediction can be completed, but the video quality deteriorates due to non-real pixel values
Solution Approach 1:
The patent introduces gradient values as an intermediary between the available reference pixels and the required out-of-bounds prediction values. These gradients act as a mediator that preserves the relationship between real pixels and their extrapolated counterparts, maintaining reliability
Solution Approach 2:
The patent changes the parameter used for prediction from simple padded values to gradient-based extrapolated values. By changing from static padding to dynamic gradient calculation, the system maintains video quality while completing predictions for out-of-bounds regions
3Measurement precision
If motion vectors are refined, then the prediction accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent applies MV refinement selectively rather than uniformly across all blocks. By refining motion vectors only where boundary conditions warrant it, the system achieves necessary accuracy without excessive computational complexity
Solution Approach 2:
The patent segments the prediction process into distinct phases: initial MV estimation, boundary condition assessment, selective refinement, and gradient-based extrapolation. This segmentation allows computational resources to be focused where they provide maximum benefit
Data Source
AI summary
A computing system receives a video bitstream that comprises a current block within a current picture and one or more syntax elements. The current block is encoded in a subblock-based motion vector (MV) refinement mode using information from a first reference block and a second reference block. The system applies a MV refinement to generate a refined first MV indicating a first location of the first reference block and a refined second MV indicating a second location of the second reference block. Subsequent to applying the MV refinement, in accordance with a determination that (i) the first location is outside of a first set of reference boundaries and/or (ii) the second location is outside of a second set of reference boundaries, the system decodes the current block using a weighted average of respective values from the first and second locations.