Chroma Intra Prediction Mode Selection with Staged RDO
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The HEVC standard's chroma intra prediction modes are inefficient due to parallel encoding of luminance and chroma components, preventing the use of the DM mode for chroma, which hinders optimal encoding efficiency.
Innovation Solution
Performing simple RDO calculations in parallel on luminance and chroma values of each pixel to determine candidate modes, followed by full RDO calculations to select target modes, allowing for the DM mode to be used during chroma encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If luminance and chroma are encoded in parallel, then encoding speed is improved, but the optimal luminance mode cannot be acquired in time for chroma encoding
Solution Approach 1:
The patent performs simple RDO calculations for luminance modes in advance during the parallel encoding process, before the full RDO calculation is completed. This preliminary identification of candidate luminance modes allows the chroma encoding to proceed with mode selection information already prepared, resolving the timing conflict between parallel encoding and optimal mode acquisition.
Solution Approach 2:
The patent segments the RDO calculation into two distinct stages: simple RDO calculation that identifies candidate modes quickly, and full RDO calculation that determines the final optimal mode. This segmentation allows different parts of the encoding process to operate at different speeds and with different levels of computational complexity, enabling parallel processing while still achieving optimal mode selection.
2Productivity
If simple RDO calculation is used for chroma mode selection, then encoding efficiency is improved, but mode selection accuracy may be compromised
Solution Approach 1:
The patent divides mode selection into two phases: simple RDO for rapid candidate identification and full RDO for final accurate selection. The simple RDO phase filters out obviously suboptimal modes quickly, while the full RDO phase performs comprehensive evaluation on the reduced candidate set, achieving both speed and accuracy.
Solution Approach 2:
The patent applies different levels of calculation complexity to different parts of the mode selection process. Simple RDO is applied broadly to all candidate modes for initial filtering, while full RDO is applied selectively to the final candidate modes that need precise evaluation, optimizing the balance between computational effort and selection accuracy.
Data Source
AI summary
A determination method for a chroma intra prediction mode includes: performing a simple RDO calculation on a luminance value of each pixel of an image block to obtain multiple luminance candidate modes; determining multiple chroma candidate modes according to one of the luminance candidate modes and a chroma simple RDO result calculated by performing a simple RDO calculation on a chroma value of each pixel of the image block; performing a full RDO calculation of the multiple luminance candidate modes on the luminance value of each pixel in the image block to select a luminance target mode; performing a full RDO calculation of the multiple chroma candidate modes on the chroma value of each pixel of the image block to a obtain chroma full RDO result; and determining a chroma target mode according to the luminance target mode and the chroma full RDO result.