Intra Prediction Sample Filtering for Complex Video Coding Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in optimizing intra prediction processes, particularly in handling complex prediction modes and residual data, which affect compression efficiency and decoding performance.
Innovation Solution
The proposed method involves deriving an intra prediction angle parameter, calculating an inverse angle, setting a scale variable based on this parameter, and performing position-dependent intra prediction sample filtering to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex prediction modes are used in intra prediction, then prediction accuracy is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the intra prediction angle based on the aspect ratio of the current block. Different angle ranges are selected depending on whether the block is wider or taller, optimizing prediction accuracy for each orientation without requiring complex mode switching. The angle is calculated using simple arithmetic operations on block dimensions rather than complex trigonometric functions.
Solution Approach 2:
The patent segments the prediction process by dividing it into distinct cases based on block orientation (wider vs. taller blocks). Each segment uses a simplified prediction approach tailored to its specific geometry, avoiding the need to implement and evaluate all possible complex prediction modes for every block type.
2Adaptability or versatility
If more prediction modes are supported, then coding flexibility is improved, but processing overhead increases
Solution Approach 1:
The patent implements partial action by providing enhanced prediction capabilities specifically for diagonal orientations, which are the most challenging to predict accurately. Rather than uniformly increasing complexity across all prediction modes, the patent focuses computational effort on the specific angle ranges that benefit most from optimization, achieving improved coding flexibility where needed without proportionally increasing overall processing overhead.
3Measurement precision
If position-dependent filtering is applied, then prediction sample accuracy is improved, but computational load increases
Solution Approach 1:
The patent applies local quality by performing position-dependent filtering selectively on prediction samples based on their location and the block's orientation. Rather than uniformly filtering all samples with complex operations, the patent applies simplified filtering appropriate to each position and orientation, improving prediction sample accuracy locally without proportionally increasing overall computational load.
Data Source
AI summary
This disclosure relates to video coding and more particularly to techniques for performing a position-dependent intra prediction sample filtering process. In one example, a method of coding video data comprises: deriving an intra prediction angle parameter; deriving an inverse angle parameter based on the intra prediction angle parameter; for a prediction mode included in a defined set of prediction modes, setting a value of a scale variable based on the derived inverse angle; and performing a position-dependent intra prediction sample filtering process using the value of a scale variable.