Intra-Prediction Edge Filtering for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards, such as H.264/MPEG-4 AVC, struggle to efficiently compress screen content with sharp edges and mixed content, which includes both screen and camera-captured content, leading to suboptimal coding efficiency and increased bit-rate due to inaccurate prediction methods.
Innovation Solution
A method for decoding and encoding video blocks using intra-prediction, where prediction modes for neighboring blocks are determined, and prediction values for edge samples are generated by applying an intra-prediction process or a filter based on the number of neighboring blocks using intra-block copy or palette modes, with a threshold to decide between these methods, thereby improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional intra-prediction methods are used for screen content with sharp edges, then coding efficiency is reduced, but bit-rate increases
Solution Approach 1:
The patent applies different prediction methods to different regions based on content characteristics. It determines whether neighboring blocks use intra-block copy or palette modes, and applies filtering selectively to edge samples only when appropriate, rather than uniformly processing all blocks. This local adaptation improves coding efficiency for screen content while maintaining acceptable bit-rate
Solution Approach 2:
The patent changes the processing parameters dynamically based on the prediction modes of neighboring blocks. When a threshold number of neighboring blocks use intra-block copy or palette modes, the edge sample generation method switches from standard intra-prediction to filtering-based generation. This parameter change optimizes the balance between coding efficiency and bit-rate for mixed content
2Measurement precision
If intra-prediction is applied to generate prediction values for edge samples, then prediction accuracy improves, but residual data increases
Solution Approach 1:
The patent applies filtering only to edge samples rather than all prediction samples, and only when the threshold condition is met. This partial application of filtering reduces the amount of residual data that needs to be encoded, while still improving prediction accuracy for the critical edge regions that contain sharp transitions in screen content
3Productivity
If filtering is applied to reference samples to generate prediction values, then coding efficiency improves, but prediction accuracy for sharp edges decreases
Solution Approach 1:
The patent applies filtering selectively only to edge samples rather than all prediction samples. Edge samples benefit from filtering when neighboring blocks indicate screen content, while interior samples use standard intra-prediction to maintain accuracy for sharp edges. This local differentiation resolves the contradiction between coding efficiency and prediction accuracy
Data Source
AI summary
A method of decoding a current block encoded using intra-prediction includes determining prediction modes for coding blocks neighboring the current block. The method generates prediction values for edge samples of the decoded block from intra-prediction reference samples of the neighboring coding blocks by applying an intra-prediction process to the intra-prediction encoded current block if a number of the neighboring coding blocks determined to one of use intra-block copy prediction mode and palette mode is greater than or equal to a predetermined threshold. Alternatively the method generates the prediction values for the edge samples of the decoded block from intra-prediction reference samples of the neighboring coding blocks by applying a filter, preferably an intra-boundary filter, to reference samples of the neighboring blocks. The current block is then decoded based on the prediction values.