Adaptive Intra Prediction Filtering for Video Coding Block Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codecs face challenges in providing continuous prediction sample surfaces across multiple prediction block boundaries, leading to inefficiencies in compression and visual quality, particularly due to discontinuities along block borders during directional intra prediction.
Innovation Solution
An apparatus and method that adaptively apply prediction sample filtering based on the directionality of spatial intra prediction, using angularity information to select appropriate filtering techniques and coefficients, ensuring continuous prediction surfaces along block boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If directional intra prediction is applied to predict pixel values in image blocks, then compression efficiency is improved, but discontinuities occur along block boundaries reducing visual quality
Solution Approach 1:
The patent applies different filtering operations to different regions of the prediction block based on their spatial location and relationship to block boundaries. Specifically, it applies filtering to prediction samples at certain distances from block boundaries while leaving other regions unchanged, thereby improving continuity at boundaries without affecting the overall compression efficiency gained from directional prediction.
Solution Approach 2:
The patent introduces adaptive filtering that dynamically adjusts prediction sample values based on their position relative to block boundaries. The filtering operation modifies prediction samples at specific distances from boundaries adaptively, allowing the system to maintain compression efficiency while improving visual quality where needed.
2Manufacturing precision
If prediction samples are filtered to improve continuity at block boundaries, then visual quality is improved, but processing complexity increases
Solution Approach 1:
The filtering operation is applied selectively only to prediction samples at specific distances from block boundaries, rather than filtering all prediction samples uniformly. This localized approach improves continuity at boundaries while minimizing the overall processing complexity by limiting filtering operations to only the necessary regions.
3Productivity
If adaptive filtering is applied based on prediction direction, then compression efficiency is improved, but the complexity of determining filter coefficients increases
Solution Approach 1:
The patent pre-establishes multiple filter coefficient sets corresponding to different prediction directions before the actual filtering process. During encoding/decoding, the system simply selects the appropriate pre-computed filter coefficients based on the prediction direction, avoiding the need to calculate complex coefficients in real-time while maintaining high compression efficiency.
Solution Approach 2:
The patent changes the parameter of filter coefficients based on prediction direction to adapt the filtering operation to different encoding scenarios. By selecting different coefficient sets according to the prediction direction, the system optimizes compression efficiency for each direction while managing complexity through predefined parameter configurations.
Data Source
AI summary
There is disclosed apparatuses, methods and computer programs for video coding. The filtering process to be applied to intra predicted sample values is selected and adapted based on the characteristics of the sample prediction process. According to some embodiments an angularity information and pixel values for filtering by a filter are received. The angularity information is used to adapt the filter. The adapting may comprise selecting a filter algorithm and/or filter coefficients. The selected pixels are then filtered by the adapted filter.