Adaptive Blending Widths for Geometric Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods using geometric partitioning mode (GPM) face inefficiencies due to the selection of a single blending width that may not match various color distributions in image frames, leading to increased bit usage and reduced coding efficiency.
Innovation Solution
A method for determining a blending width in GPM by comparing block size with a threshold, using adjacent block template regions, and weightedly combining predicted blocks to reconstruct the block unit, thereby optimizing the selection of blending width based on cost values and block size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a single blending width is used in GPM, then the coding process is simple, but it cannot match various color distributions in image frames, leading to reduced coding efficiency
Solution Approach 1:
The patent applies dynamics by making the blending width adaptive rather than fixed. The blending width is dynamically selected from multiple candidate widths based on the actual color distribution characteristics of each block unit, allowing the coding process to adapt to different local regions while maintaining reasonable complexity through automated selection.
Solution Approach 2:
The patent changes the parameter of blending width from a single fixed value to multiple candidate values that can be selected based on color distribution. By introducing multiple candidate widths and selecting the most appropriate one for each block, the system improves coding efficiency without excessively complicating the overall process.
2Measurement precision
If the last one of the candidate widths is always selected for each block, then the color distribution matching is precise, but the number of bits increases, negatively impacting coding efficiency
Solution Approach 1:
The patent applies local quality by allowing different blending width selections for different block units based on their specific color distribution characteristics. Rather than uniformly selecting the last candidate width for all blocks, the system evaluates each block locally and selects the most appropriate width, achieving precise matching where needed while avoiding unnecessary bit usage in other regions.
Solution Approach 2:
The patent uses partial action by selecting from multiple candidate widths rather than always using the maximum (last) width. This allows the system to apply the full precision of the last candidate width only when necessary for specific color distributions, while using fewer bits for blocks that can be adequately represented by narrower widths.
3Measurement precision
If multiple candidate widths are evaluated for each block unit, then the blending width selection is precise, but the computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining a set of candidate blending widths that are likely to be optimal for different block types. This preliminary preparation allows the decoder to quickly evaluate and select from a manageable set of candidates rather than searching through all possible widths, reducing computational complexity while maintaining selection precision.
Data Source
AI summary
A method of decoding video data by an electronic device is provided. The electronic device receives the video data and determines a block unit from a current frame included in the video data. The electronic device determines, from a plurality of candidate lines, a partitioning line for dividing the block unit into a pair of geometric partitions using a geometric partitioning mode and, from a plurality of candidate modes, two different prediction modes of the block unit to generate two predicted blocks for the pair of the geometric partitions. The electronic device determines a blending width of the partitioning line from a plurality of candidate widths based on a block size of the block unit and weightedly combining the two predicted blocks along the partitioning line based on the blending width to reconstruct the block unit.