Cost-Ranked MPM List Construction for Lower-Bit Intra-Prediction Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing intra-prediction technologies require a large number of bits for encoding indexes of the MPM list due to the fixed inclusion of primary and secondary MPM lists, which are ordered and include multiple prediction modes.
Innovation Solution
A method to determine M first intra-prediction modes based on prediction cost, perform mode offset on angular prediction modes, and establish a most likely mode MPM list, reducing the number of bits required for encoding by optimizing the MPM list construction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the MPM list includes a primary MPM list with 6 prediction modes and a secondary MPM list with 16 prediction modes formed in a preset order, then the prediction modes are systematically organized, but a relatively large number of bits are required for encoding indexes of the MPM list
Solution Approach 1:
The patent applies dynamics by making the MPM list construction adaptive rather than static. The list is dynamically constructed based on prediction costs calculated from neighboring block modes, allowing the composition and ordering of prediction modes to vary according to actual video content characteristics, thereby optimizing encoding efficiency
Solution Approach 2:
The patent changes the parameter of prediction mode ordering from a fixed preset sequence to a cost-based sequence. By calculating prediction costs for different modes and arranging them in ascending order of cost, the system transforms the MPM list into a content-adaptive structure that reduces encoding bits while maintaining prediction accuracy
Data Source
AI summary
Embodiments of this application provide a list construction method and a terminal. The list construction method includes: determining M first intra-prediction modes based on a prediction cost corresponding to each of N first intra-prediction modes; where the M first intra-prediction modes are at least part of the N first intra-prediction modes; performing mode offset on at least part of angular prediction modes in the M first intra-prediction modes based on an order of the M first intra-prediction modes to obtain K second intra-prediction modes; and establishing a most likely mode MPM list based on the M first intra-prediction modes and the K second intra-prediction modes; where the MPM list includes the M first intra-prediction modes and the K second intra-prediction modes.