Intra-frame Chroma Prediction Mode Selection in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video coding technologies, such as HEVC, face low calculation efficiency due to the need to calculate coding cost values for each candidate intra-frame prediction mode during the prediction mode selection process, leading to increased computational complexity.
Innovation Solution
A prediction mode selection method that reduces calculation by performing a primary selection using prediction residual costs and a secondary selection using coding cost values, effectively narrowing down the candidates and improving coding efficiency by calculating difference evaluation values, which are 1/6-1/4 of the coding cost values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If coding cost value calculation is performed for each candidate intra-frame prediction mode, then accurate mode selection is achieved, but calculation complexity increases significantly
Solution Approach 1:
The patent segments the prediction mode selection process into two distinct stages: a first stage that performs comprehensive coding cost value calculations for all candidate modes to ensure accuracy, and a second stage that uses a simplified calculation method for routine selections. This segmentation allows the system to maintain high selection accuracy while reducing overall computational complexity by avoiding full-cost calculations in every scenario.
Solution Approach 2:
The patent performs preliminary classification of prediction modes based on their characteristics and likelihood of being optimal. By pre-identifying and prioritizing candidate modes before performing expensive coding cost calculations, the system reduces the number of full calculations needed, thereby lowering computational complexity while maintaining selection accuracy through targeted evaluation of the most promising candidates.
2Measurement precision
If coding cost value calculation is performed for each candidate intra-frame prediction mode, then optimal mode is selected, but calculation time increases
Solution Approach 1:
The patent divides the mode selection process into time-efficient segments: rapid preliminary assessment of candidate modes followed by detailed coding cost calculation only for the most promising candidates. This segmentation reduces total calculation time by avoiding comprehensive calculations for all modes while preserving selection accuracy through focused evaluation.
Solution Approach 2:
The patent applies partial action by performing complete coding cost calculations only for a subset of candidate modes rather than all candidates. By calculating difference evaluation values for selected candidates and using these partial results to determine optimal modes, the system achieves acceptable accuracy with significantly reduced calculation time.
3Productivity
If comprehensive coding cost value calculation is performed, then coding efficiency is improved, but calculation amount increases
Solution Approach 1:
The patent segments the calculation workload into essential difference evaluation calculations that directly impact coding efficiency and supplementary full coding cost calculations. By separating these calculation types and prioritizing the efficient difference evaluations, the system maintains coding efficiency improvements while reducing the total quantity of calculations required.
Solution Approach 2:
The patent uses partial action by calculating difference evaluation values for a limited set of candidate modes rather than performing exhaustive coding cost calculations for all modes. This partial calculation approach achieves sufficient coding efficiency improvement without the prohibitive computational burden of comprehensive evaluation, optimizing the balance between productivity gain and calculation resource consumption.
Data Source
Figure 1~2(D)
Figure 3
Figure 4
AI summary
A prediction mode selection method, apparatus and system, which falls within the field of video coding. The method comprises: screening a first group of candidate intra-frame chroma prediction modes of a current prediction unit in N intra-frame prediction modes; respectively performing prediction residual cost calculation on the first group of candidate intra-frame chroma prediction modes to obtain a difference evaluation value; screening a second group of candidate intra-frame chroma prediction modes according to the difference evaluation value; respectively performing coding cost calculation on the second group of candidate intra-frame chroma prediction modes to obtain a coding cost value; and determining an intra-frame chroma prediction mode with the minimum coding cost as a final intra-frame chroma prediction mode according to the coding cost value. Since the calculation amount of a difference evaluation value is 1/6-1/4 of the calculation amount of a coding cost value, the effects of effectively reducing the calculation amount and improving the coding efficiency are achieved.