Intra Prediction Mode Selection via Cost Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
H.264 intra prediction requires a significant amount of computing time to determine the optimum prediction mode due to the large number of cost calculations needed across various prediction methods, which is inefficient, especially as display apparatuses become larger.
Innovation Solution
The method involves performing cost calculations in order from larger encoding sizes, selecting prediction modes based on previous results, and comparing minimum costs to reduce the number of times cost calculation is performed, allowing for efficient intra prediction by limiting search target modes in each prediction method.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple prediction methods with various patterns are used to determine the optimum prediction mode, then the accuracy of prediction is improved, but the computing time and complexity increase significantly
Solution Approach 1:
The macroblock is divided into multiple subblocks, and intra prediction is performed separately for each subblock. This segmentation allows the prediction process to be broken down into smaller, more manageable units, reducing the overall computing complexity while maintaining prediction accuracy for each segment
Solution Approach 2:
Instead of performing all nine prediction patterns for every subblock, the invention selectively applies prediction patterns based on the subblock size and characteristics. For 16x16 subblocks, only four prediction patterns are used, while 8x8 and 4x4 subblocks use nine patterns. This partial action approach reduces unnecessary computations while maintaining adequate prediction accuracy
2Measurement precision
If multiple prediction methods with various patterns are used to determine the optimum prediction mode, then the accuracy of prediction is improved, but the device complexity increases
Solution Approach 1:
The prediction process is segmented by subblock size (16x16, 8x8, 4x4), with different numbers of prediction patterns applied to each segment. This segmentation simplifies the overall device complexity by avoiding the need to implement all nine patterns for all subblock types, while still maintaining high prediction accuracy where needed
Solution Approach 2:
The invention implements a partial set of prediction patterns (four patterns) for 16x16 subblocks, which reduces the processing complexity compared to implementing all nine patterns. This partial action is sufficient for larger subblocks while more comprehensive patterns are used for smaller subblocks where higher precision is needed
Data Source
AI summary
Search is performed on Intra 16 to obtain a prediction mode leading to a minimum cost, and the minimum cost in Intra 16 and a corresponding prediction mode are stored. Search is performed on Intra 8 to obtain a prediction mode leading to a minimum cost, and then a relationship of magnitude between the stored minimum cost in Intra 16 and the minimum cost in Intra 8 is judged. After that, the minimum cost in Intra 8 and a corresponding prediction mode are stored, and search is performed on Intra 4 to obtain a prediction mode leading to a minimum cost. A relationship of magnitude between cost_intra and the minimum cost in Intra 4 is judged, and Intra 4 is determined as an optimum prediction mode in a case where Intra 4 is smaller.