Video Subblock IntraTMP for Complete Template Availability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently utilizing intra template matching prediction at the subblock level due to issues such as incomplete template availability for subblocks, leading to suboptimal prediction accuracy.
Innovation Solution
The proposed method involves partitioning a current block into subblocks and determining templates for these subblocks based on reconstructed samples and portions of the current block, using intra template matching prediction (IntraTMP) to ensure complete template availability for accurate prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra template matching prediction is applied at the subblock level, then prediction accuracy is improved, but template availability becomes incomplete for some subblocks
Solution Approach 1:
The patent applies nesting by constructing subblock templates from multiple sources: when a subblock lacks sufficient adjacent samples, it borrows samples from parent block templates and previously reconstructed neighboring subblocks. This nested structure allows templates to be built hierarchically from available data at different levels (current block, parent block, adjacent subblocks), ensuring complete template availability for all subblocks while maintaining prediction accuracy
Solution Approach 2:
The patent introduces reconstructed samples from previously processed subblocks as intermediary data to bridge the gap between unavailable original templates and required prediction templates. These intermediate reconstructed samples serve as substitutes that enable template construction for subblocks that would otherwise lack sufficient reference data
2Measurement precision
If the current block is partitioned into multiple subblocks for intra template matching prediction, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent segments the current block into multiple subblocks to enable finer-grained template matching prediction. Each subblock can independently select its template from available sources (adjacent samples, parent block template, reconstructed neighboring subblocks), allowing localized optimization while maintaining overall efficiency through systematic template construction rules
Solution Approach 2:
The patent performs preliminary reconstruction of subblocks in a specific order (e.g., raster scan order) so that reconstructed samples from earlier subblocks are available for template construction of subsequent subblocks. This preliminary action ensures that when processing each subblock, the necessary reference data from previously reconstructed subblocks is already in place, eliminating the need for complex backward references
Data Source
AI summary
An apparatus of video decoding is provided. The apparatus includes processing circuitry. The processing circuitry is configured to receive a video bitstream including coded information of a current block and a template of the current block in a current picture. The template of the current block includes samples adjacent to the current block, and the current block includes a first subblock and a second subblock adjacent to the first subblock. The processing circuitry is configured to determine a template of the second subblock based on at least one of (i) the template of the current block and (ii) reconstructed samples of the first subblock. The processing circuitry is configured to reconstruct the second subblock based on the determined template according to intra template matching prediction (intraTMP).