Video Intra-Prediction Using Multi-Line Reference Sample Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding/decoding technologies face challenges in efficiently performing intra-prediction for high-resolution video signals, particularly in replacing unavailable reference samples and calculating average values based on multiple reference lines.
Innovation Solution
A method and apparatus that derive multiple reference sample lines for a current block, select an available reference sample line, and replace unavailable samples with the nearest available sample, using the selected line for intra-prediction, and calculate the prediction sample based on the average value of reference samples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple reference sample lines are derived for intra-prediction in high-resolution video, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The reference samples are segmented into multiple reference lines (first reference line, second reference line, etc.), each containing samples at different distances from the current block. This segmentation allows selective use of reference lines based on prediction mode and block characteristics, improving accuracy while managing complexity through organized structure.
Solution Approach 2:
The patent dynamically selects which reference line to use based on the intra-prediction mode and block characteristics. Different reference lines are activated conditionally - for example, using the first reference line for certain prediction modes and the second reference line for others. This dynamic selection optimizes prediction accuracy while avoiding unnecessary processing of all reference lines in all cases.
2Reliability
If unavailable reference samples are replaced with available samples, then prediction completeness is improved, but prediction accuracy may deteriorate
Solution Approach 1:
When reference samples are unavailable (e.g., at block boundaries or in unavailable regions), the patent uses available reference samples from the same or adjacent reference lines as intermediaries to substitute for the missing samples. This substitution ensures the prediction process can continue complete ly while minimizing accuracy loss by using the nearest available samples.
Solution Approach 2:
The patent performs preliminary checks to identify unavailable reference samples before the prediction process, and prepares replacement values in advance by selecting appropriate available samples from the reference lines. This preliminary action ensures that when prediction is performed, all necessary reference samples are ready, maintaining both completeness and accuracy.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A method for processing a video according to the present invention may comprise: generating a plurality of Most Probable Mode (MPM) candidates; determining whether there is an MPM candidate identical to an intra-prediction mode of a current block among the plurality of MPM candidates; obtaining the intra-prediction mode of the current block, based on a result of the determining; and performing an intra-prediction for the current block, based on the intra-prediction mode of the current block.