Intra Prediction Mode Selection Using Progressive Cost Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding systems are computationally intensive and time-consuming when attempting to find the best intra prediction mode for each block in video coding, particularly in H.265/HEVC, due to the need to evaluate all possible modes at every possible block size, which can lead to increased processing time without a suitable tradeoff between computational complexity and decoded video quality.
Innovation Solution
A method that progressively narrows down potential intra prediction modes over multiple stages by dividing pixels into groups and calculating cost values and threshold maximum costs at each stage, ultimately selecting the mode with the lowest cost value for each block, while also considering heterogeneity indices for contiguous blocks to finalize modes and reduce unnecessary processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all possible intra prediction modes are evaluated at every possible block size, then decoded video quality is improved, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent segments the pixel set into multiple groups (e.g., first group, second group, third group, fourth group) and processes each group separately through multiple stages. This segmentation allows the encoder to evaluate prediction modes in a divided manner, reducing the computational burden of evaluating all modes for all pixels simultaneously while maintaining accurate quality assessment through progressive refinement across stages.
Solution Approach 2:
The patent performs preliminary evaluation of intra prediction modes by calculating cost values for groups of pixels in early stages before finalizing the selection. By computing preliminary cost values for multiple groups and comparing them across stages, the encoder can eliminate suboptimal modes early in the process, avoiding the need to perform exhaustive evaluations of all modes for all pixels, thus reducing overall computational complexity while preserving quality.
2Measurement precision
If all possible intra prediction modes are evaluated at every possible block size, then decoded video quality is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the pixel set into multiple groups and processes them through multiple stages, evaluating cost values for each group separately. This segmentation enables parallel or sequential processing of different pixel groups, reducing the overall processing time compared to evaluating all pixels for all modes simultaneously, while still achieving accurate mode selection through progressive refinement across stages.
Solution Approach 2:
The patent performs preliminary cost value calculations for multiple pixel groups in early stages to identify promising prediction modes before final selection. By computing preliminary metrics for groups of pixels rather than waiting to evaluate complete blocks, the encoder can eliminate suboptimal modes early, significantly reducing the time required for exhaustive evaluation while maintaining quality through multi-stage refinement.
3Device complexity
If the number of intra prediction mode candidates is reduced, then computational complexity is decreased, but prediction accuracy may be compromised
Solution Approach 1:
The patent performs preliminary evaluation of all intra prediction modes by calculating cost values for multiple pixel groups in early stages. This preliminary action identifies the most promising modes based on group-level metrics, allowing the encoder to limit subsequent detailed evaluation to only these candidate modes. This approach reduces computational complexity by avoiding exhaustive evaluation of all modes while preserving prediction accuracy through thorough preliminary assessment of candidates.
Solution Approach 2:
The patent segments the evaluation process into multiple stages, where early stages evaluate all modes using simplified cost calculations on pixel groups, and later stages perform more detailed evaluation on a reduced set of candidate modes. This segmented approach ensures that no potentially accurate mode is eliminated prematurely by simplified metrics, while still achieving computational efficiency by limiting detailed evaluation to promising candidates only.
Data Source
AI summary
A method is provided for encoding a digital video with a video encoder that allows intra prediction of mode selection. The method includes dividing pixels from each intra prediction block into four groups based on phases of the pixels, progressively narrowing down a list of intra prediction modes over a series of stages based on a comparison of a cost value for each intra prediction mode at that stage against a threshold value, tentatively assigning the remaining intra prediction with the lowest cost value after the final stage to the block, calculating a heterogeneity index for each group of contiguous intra prediction blocks based on the tentative intra prediction mode selected for each block after the final stage, and finalizing the tentative intra prediction mode for each block if the group has a heterogeneity index above a predetermined threshold value.