Recursive Intra-Region Coding With Implicit Inter-Mode Derivation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods face inefficiencies in signaling prediction modes for the last coding block, leading to increased overhead and processing requirements.
Innovation Solution
Implicitly derive the prediction mode of the last coding block based on a signaled flag indicating that at least one coding block in the coding region is encoded in a specific mode, without requiring additional indicators for the last block.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If explicit prediction mode indicators are signaled for every coding block in the bitstream, then the decoder can accurately reconstruct each block's prediction mode, but the signaling overhead increases and compression efficiency decreases
Solution Approach 1:
The patent extracts the prediction mode information for the last coding block from the explicit signaling process by inferring it from the coding region's overall prediction mode indicator. This removes the redundant overhead of separately signaling the last block's mode while preserving the ability to accurately reconstruct it during decoding.
Solution Approach 2:
The last coding block's prediction mode is determined through self-inference based on the coding region's predicted mode indicator and the processing order of blocks. The decoder uses the already-signaled region-level indicator to automatically derive the last block's mode without requiring additional explicit signaling, making the system self-sufficient.
2Reliability
If prediction mode indicators are signaled for all coding blocks, then complete prediction information is available, but the processing complexity and computational requirements increase
Solution Approach 1:
The patent extracts the unnecessary processing step of explicitly handling prediction mode indicators for the last coding block. By inferring the mode from the region-level indicator, the decoder eliminates redundant processing operations while maintaining complete prediction information for all blocks.
Solution Approach 2:
Instead of signaling prediction modes for all blocks (excessive action), the patent applies partial signaling only to non-last blocks within the coding region. The last block's mode is derived partially from the region-level indicator, reducing overall processing complexity while maintaining information completeness.
3Measurement precision
If the bitstream includes explicit prediction mode indicators for the last block, then decoding accuracy is maintained, but the bitstream length and transmission bandwidth requirements increase
Solution Approach 1:
The patent removes the redundant prediction mode indicator for the last coding block from the bitstream. The decoder infers this information from the coding region's predicted mode indicator, maintaining decoding accuracy while reducing the total bitstream length and transmission requirements.
Solution Approach 2:
Instead of explicitly signaling the last block's prediction mode (traditional approach), the patent inverts the approach by having the decoder infer the mode from the region-level indicator. This inversion reduces bitstream overhead while preserving decoding precision through logical derivation.
Data Source
AI summary
An example method of video coding includes receiving a video bitstream including a plurality of coding blocks; identifying, based on a first indicator in the video bitstream, a coding region that comprises two or more blocks of the plurality of coding blocks, where at least one block in the coding region is encoded in an intra prediction mode and at least one block in the coding region is encoded in a non-intra prediction mode. For a last block in the coding region, when all preceding blocks in the coding region are encoded in the intra prediction mode, the method includes inferring a prediction mode of the last block as being an inter prediction mode without receiving an inter prediction mode indicator for the last block; and reconstructing the last block using the inferred prediction mode.