IBC-Constrained Sign Derivation for Block Vector Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently encoding and decoding video sequences, particularly in handling screen content with complex structures, leading to increased computational complexity and bitrate requirements.
Innovation Solution
The implementation of intra block copy (IBC) reference region constraints on block vector differences (BVD) signaling, which conditionally signals the sign of directional components in the bitstream, optimizing the encoding and decoding process for screen content by reducing redundancy and improving coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video coding techniques are used, then encoding and decoding can be performed, but computational complexity and bitrate requirements increase significantly for screen content
Solution Approach 1:
The patent changes the parameter representation by signaling only the sign of block vector difference components rather than full precision values. This parameter change reduces computational complexity while maintaining coding efficiency for screen content where block vectors follow predictable patterns.
Solution Approach 2:
The patent segments the block vector difference signaling into separate components (sign and magnitude), where the sign is signaled separately based on reference region constraints. This segmentation allows for more efficient encoding by exploiting the structured nature of screen content block vectors.
2Productivity
If conventional video coding techniques are used, then encoding and decoding can be performed, but bitrate requirements increase significantly
Solution Approach 1:
The patent extracts only the necessary sign information from the block vector difference components, removing redundant magnitude information that can be inferred from reference region constraints. This extraction significantly reduces the bitrate required for encoding screen content while maintaining decoding accuracy.
3Measurement precision
If block vector differences are fully signaled, then decoding accuracy is maintained, but redundancy in the bitstream increases
Solution Approach 1:
Instead of signaling the complete block vector difference and then removing redundancy, the patent inverts the approach by directly signaling only the essential sign information and deriving the magnitude from the reference region constraints. This inversion eliminates redundancy at the source while maintaining decoding accuracy.
Data Source
AI summary
A decoder receives, in a bitstream for a block, an indication of a block vector predictor (BVP) and indication of magnitudes associated with a first and a second component of a block vector difference (BVD). BV candidates are determined based on a sum of the BVP and a different one of BVD candidates. Each of the BVD candidates includes: the magnitudes associated with the first and second components, and a different combination of signs for the first and second components. The decoder determines a sign for at least one of the first and second components of the BVD based on validities of the BV candidates indicating only two or three of the BV candidates being valid. Each of the validities for a respective BV candidate of the BV candidates is based on whether the BV candidate indicates a reference block located within a reference region for the block.