Intra-Prediction Filter Selection for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face inefficiencies in intra-prediction methods, particularly in determining the optimal filtering process for neighboring samples, which affects coding cost and image quality.
Innovation Solution
A method that selects between a default filter and an N-tap filter based on the intra-prediction mode and block size, transmitting a flag when necessary, to generate filtered neighboring samples for improved prediction in video coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a default filter is used for filtering neighboring samples, then the device complexity is reduced, but the image quality deteriorates
Solution Approach 1:
The patent implements dynamic filter selection where the filtering process adapts based on block characteristics. The system switches between default filter and more complex filtering methods depending on the specific block being processed, allowing complexity to vary dynamically rather than being fixed, thus resolving the contradiction between device complexity and image quality.
Solution Approach 2:
The patent applies different filtering strategies to different blocks based on their local characteristics. By analyzing block-specific features and applying appropriate filter types (default or more sophisticated filtering) only where needed, the system achieves high image quality in critical areas while maintaining low complexity in other areas, resolving the contradiction between overall complexity and overall quality.
2Manufacturing precision
If a more sophisticated filtering process is applied to neighboring samples, then the image quality is improved, but the coding cost increases
Solution Approach 1:
The patent changes the parameter of filter selection based on block characteristics. By adjusting which filter type is applied (default vs. sophisticated filtering) according to block-specific parameters, the system optimizes the balance between image quality improvement and coding cost, applying complex filtering only when it provides sufficient benefit to justify the additional cost.
Solution Approach 2:
The patent applies sophisticated filtering only partially - specifically to blocks where it is most beneficial - rather than uniformly to all blocks. This selective application of advanced filtering techniques achieves significant quality improvement while avoiding the excessive coding cost that would result from applying the same sophisticated filtering to every block, thus resolving the contradiction between quality and efficiency.
Data Source
AI summary
Aspects of the disclosure include a method for video coding. The method includes receiving input data associated with a current block in a current image frame of video data, where the current block is coded by intra-prediction or to be coded by intra-prediction. The method also includes determining an intra-prediction mode of the current block, selecting one of a plurality of filters including at least a default filter and an N-tap filter, and generating filtered neighboring samples by filtering neighboring samples adjacent to the current block using the selected filter, where N is a positive integer different from 3. Moreover, the method includes encoding or decoding the current block by predicting the current block based on the filtered neighboring samples and the intra-prediction mode.