Intra Block Copy with Template Matching and GPM for Efficient Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently compressing video data while maintaining quality, particularly in handling motion vectors and spatial-temporal correlations in video frames, leading to increased data requirements.
Innovation Solution
Implementing Intra Block Copy (IBC) mode combined with Geometric Partitioning Mode (GPM) and Template Matching (TM) to enhance prediction accuracy and reduce data requirements by utilizing spatial and temporal correlations, along with refined merge candidates and block vector differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If Intra Block Copy mode is used for video compression, then coding efficiency is improved, but the complexity of handling motion vectors and spatial-temporal correlations increases
Solution Approach 1:
The current block is divided into multiple sub-blocks, and different prediction modes (IBC and intra prediction) are applied to different sub-blocks. This segmentation allows the system to handle complex motion vectors only where necessary while using simpler intra prediction for other regions, thereby improving coding efficiency without uniformly increasing complexity across the entire block.
Solution Approach 2:
Instead of applying IBC mode with full motion vector handling to the entire block, the patent applies it only to specific sub-blocks where it provides benefit. The decision to use IBC or intra prediction for each sub-block is made based on local characteristics, allowing partial application of the more complex IBC mechanism only where needed, thus improving efficiency while controlling overall complexity.
2Measurement precision
If multiple merge candidates including Template Matching refined candidates are used, then prediction accuracy is improved, but the amount of data required increases
Solution Approach 1:
The patent extracts only the essential motion vector information from Template Matching and other advanced techniques, rather than transmitting full candidate lists. By taking out only the necessary block vector differences and essential prediction parameters, the system achieves high prediction accuracy while minimizing the amount of data that needs to be encoded and transmitted.
Solution Approach 2:
The patent changes the representation parameters from full motion vectors to block vector differences (BVD) relative to a reference candidate. This parameter transformation reduces the data required while maintaining prediction accuracy, as BVD values are typically smaller and can be more efficiently encoded than complete motion vector pairs.
3Productivity
If Intra Block Copy mode combined with Geometric Partitioning Mode is used, then video compression is enhanced, but the complexity of processing syntax elements and prediction modes increases
Solution Approach 1:
The patent merges IBC mode with Geometric Partitioning Mode (GPM) by applying geometric partitioning to the reference block selection process in IBC. This combination allows the system to handle complex spatial correlations through GPM's geometric regions while using IBC for temporal prediction, achieving enhanced compression by addressing both spatial and temporal dimensions simultaneously.
Solution Approach 2:
The patent adds a geometric dimension to the traditional IBC approach by introducing GPM. Instead of only considering temporal prediction from previous frames, the system now also incorporates geometric spatial partitioning within the current frame, creating a two-dimensional prediction approach (temporal + spatial geometric) that enhances compression efficiency.
Data Source
AI summary
Methods for video decoding and encoding, apparatuses and non-transitory computer-readable storage media thereof are provided. In one method for video decoding, a decoder may obtain a current coding unit (CU) that is coded based on Intra Block Copy (IBC) mode combined with Geometric Partitioning Mode (GPM). Additionally, the decoder may obtain from at least one of regular merge candidates, Template Matching (TM) refined merge candidates, or merge candidates with block vector difference (MBVD candidates), a first IBC merge prediction and a second IBC merge prediction. Further, the decoder may obtain a prediction for the current CU based on the first IBC merge prediction and the second IBC merge prediction.