Region-Based Intra Mode Derivation for Coding Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in deriving intra prediction modes, particularly when using unconventional modes, leading to suboptimal compression and decoding performance.
Innovation Solution
The method involves deriving a second intra prediction mode from a pre-defined region that includes current and previously coded coding tree units, updating this region, and using it to select transforms and construct most probable mode lists, while handling candidate blocks within and outside this region to enhance intra prediction efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If unconventional intra prediction modes (IBC, IntraTMP, MIP, Palette) are used for coding blocks, then coding flexibility and adaptability to different content types are improved, but the complexity of deriving prediction modes and selecting transforms increases
Solution Approach 1:
The patent introduces an intermediary mechanism by deriving a second intra prediction mode from conventional modes (DC, Planar, angular) based on the first unconventional intra prediction mode. This derived mode serves as a mediator that simplifies transform selection and MPM list construction, reducing the complexity burden while preserving the adaptability benefits of unconventional modes.
Solution Approach 2:
The patent changes the parameter representation by mapping unconventional intra prediction modes to derived conventional modes. Instead of directly handling complex unconventional modes for all operations, the system transforms them into equivalent conventional mode parameters that are simpler to process for transform selection and prediction mode derivation.
2Measurement precision
If a pre-defined region including current and previously coded CTUs is used for candidate block selection, then prediction accuracy is improved, but the computational overhead and processing time increase
Solution Approach 1:
The patent applies preliminary action by pre-defining the region including current and previously coded CTUs before the actual prediction process. This pre-established region framework allows for efficient candidate block selection without repeated region calculations, reducing processing time while maintaining prediction accuracy.
Solution Approach 2:
The patent segments the picture into CTU (Coding Tree Unit) regions and defines specific pre-defined regions for candidate block search. By dividing the search space into manageable CTU-based segments rather than searching the entire picture, the system improves prediction accuracy within relevant regions while reducing overall computational overhead.
3Productivity
If derived second intra prediction modes are used for transform selection and MPM list construction, then compression efficiency is improved, but the encoding/decoding complexity increases
Solution Approach 1:
The patent implements feedback by using the derived second intra prediction mode to inform both transform selection and MPM list construction. The derived mode provides feedback information that guides subsequent encoding/decoding operations, improving compression efficiency through better-adapted transforms and prediction mode lists while the feedback mechanism itself is designed to be computationally efficient.
Data Source
AI summary
Aspects of the disclosure include methods and an apparatus for video coding. The apparatus includes processing circuitry that receives coded information of a current block in a current picture. The current block is coded with a first intra prediction mode that is different from a plurality of intra prediction modes including a DC mode, a Planar mode, and angular intra prediction modes. A candidate block in the current picture coded by an intra prediction mode different from the plurality of intra prediction modes is selected. If the candidate block is located within a pre-defined region, a second intra prediction mode is derived from the plurality of intra prediction modes for the candidate block. The second intra prediction mode is associated with the current block and is used to select a transform for the current block and/or to construct a most probable mode (MPM) list for another block.