MPM List Composition Using Non-Adjacent Block Offsets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques, such as H.264/AVC, HEVC, and VVC, face challenges in achieving higher coding efficiency and improved image quality due to increasing image size, resolution, and frame rate, particularly in efficiently composing the Most Probable Mode (MPM) list for intra prediction.
Innovation Solution
A method and apparatus for video coding that derive intra-prediction modes for neighboring blocks to compose a Most Probable Mode (MPM) list, using history-based composition, default modes, and neighbor prediction modes generated by applying offsets to directional modes, to improve video coding efficiency and enhance video quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the MPM list is composed using only adjacent neighboring blocks, then the composition process is simple, but the coding efficiency is limited due to insufficient prediction mode diversity
Solution Approach 1:
The patent segments the MPM list composition process into multiple stages: first composing the list using adjacent neighboring blocks (left and above blocks), then separately deriving and adding prediction modes from non-adjacent neighboring blocks (above-right and below-left blocks). This segmentation allows the system to maintain the simplicity of using adjacent blocks while enhancing coding efficiency through the additional diversity provided by non-adjacent blocks.
Solution Approach 2:
The patent extends the MPM list composition from a single dimension (adjacent neighboring blocks only) to multiple dimensions by incorporating non-adjacent neighboring blocks in different spatial positions. This dimensional expansion includes deriving modes from above-right and below-left blocks, as well as applying offset operations to directional modes, thereby enriching the prediction mode diversity without overwhelming complexity.
2Adaptability or versatility
If more neighboring blocks are used to compose the MPM list, then the prediction mode diversity increases, but the computational complexity increases
Solution Approach 1:
The patent applies partial action by selectively using only certain non-adjacent neighboring blocks (above-right and below-left) rather than all possible neighboring blocks. Additionally, it uses offset operations on directional modes as a supplementary method rather than exhaustively deriving all possible prediction modes. This partial approach provides sufficient prediction mode diversity while controlling computational complexity.
Solution Approach 2:
The patent performs preliminary composition of the MPM list using adjacent neighboring blocks before adding modes from non-adjacent blocks. This preliminary action establishes a baseline MPM list that can be used if computational resources are limited, while allowing optional enhancement with additional modes from non-adjacent blocks when more computational resources are available.
3Productivity
If non-adjacent blocks are included in MPM list composition, then coding efficiency improves, but the derivation process becomes more complex
Solution Approach 1:
The patent uses offset operations as an intermediary method to derive prediction modes from non-adjacent neighboring blocks. Instead of directly and complexly deriving modes from blocks at arbitrary positions, the system applies offset operations to existing directional modes, which serves as a simplified intermediary step that bridges the gap between available block data and required prediction modes.
Solution Approach 2:
The patent changes parameters by applying offset values to directional modes when deriving prediction modes from non-adjacent neighboring blocks. This parameter change approach allows the system to generate diverse prediction modes through simple arithmetic operations on mode indices, reducing the difficulty of mode derivation while maintaining coding efficiency.
Data Source
AI summary
A method is disclosed for composing a MPM list in intra prediction. A video coding method and an apparatus derive intra-prediction modes for neighboring blocks that are non-usable for composing a Most Probable Mode list (MPM list) to organize the derived intra-prediction modes into an MPM list. The video coding method and the apparatus organize, based on history-based MPM list composition, the intra-prediction modes of non-adjacent blocks into a MPM list.