Video Block Prediction Using Shared ITMP and IBC Block Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression systems, such as HEVC and VVC, do not effectively integrate intra-template-matching (ITMP) and intra block copy (IBC) prediction modes, leading to suboptimal compression performance when both modes are enabled in a video picture.
Innovation Solution
The method involves deriving a block vector between a current block and a best prediction block using ITMP, and reusing this block vector for IBC prediction modes, allowing interaction and propagation of block vectors between neighboring blocks predicted by either ITMP or IBC, thereby optimizing block vector signaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If both ITMP and IBC prediction modes are enabled in a video picture, then prediction accuracy is improved, but the complexity of integrating and signaling block vector information increases
Solution Approach 1:
The patent merges ITMP and IBC prediction modes into a unified processing framework where block vectors derived from ITMP are reused for IBC prediction. This integration allows the system to leverage the accuracy of template matching while avoiding the complexity of completely separate processing paths, achieving both prediction accuracy improvement and signaling efficiency.
Solution Approach 2:
The block vector derived through ITMP serves multiple functions: it is first used for ITMP prediction, then reused for IBC prediction mode as well. This multi-functionality reduces the need for separate block vector signaling in IBC mode, simplifying the overall system complexity while maintaining the benefits of both prediction modes.
2Reliability
If block vector information is fully signaled for IBC predicted blocks, then prediction reliability is improved, but the bitrate cost increases
Solution Approach 1:
The patent recovers and reuses block vector information that would otherwise be redundant. The block vector derived from ITMP is discarded as separate signaling in IBC mode and instead recovered through reuse, eliminating the need for additional bitrate expenditure while maintaining prediction reliability through the shared block vector information.
Data Source
AI summary
A method of predicting a block of a video picture, includes: deriving a first prediction block from an intra-template-matching prediction mode determining the first prediction block by minimizing a cost function calculated between sample values of a L-shaped template of the block of the video picture and a L-shaped template of a reconstructed block of the video picture; and determining a first block vector as a displacement between the first prediction block and the block of the video picture, the first block vector identifying the first prediction block as a prediction block candidate of the block of the video picture.