Adaptive Intra Prediction Filtering for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques, such as HEVC, are insufficient for achieving high video quality and compression efficiency, especially in high-resolution and high-refresh-rate applications like virtual reality, due to limitations in intra prediction modes and reference sample smoothing, leading to potential sample errors and increased computational complexity.
Innovation Solution
The method involves adaptive filtering for intra prediction using a multiple reference line (MRL) approach, where a filter index is generated based on an input index to determine whether to perform filtering or directly output the input sample, reducing computational complexity and improving compression efficiency for super high-resolution videos.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reference smoothing filtering (RSF) is applied to all intra prediction modes, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent changes the parameter of filter application from universal to selective by introducing an index-based control mechanism. The filter is applied only when the index indicates a single reference line is used, rather than applying it to all intra prediction modes uniformly.
Solution Approach 2:
The patent implements dynamic filter selection where the application of reference smoothing filtering is determined at runtime based on the index value. The system adapts its behavior (applying or skipping the filter) according to the specific prediction mode being used, rather than using a static universal approach.
2Adaptability or versatility
If multiple reference lines are used for intra prediction, then prediction flexibility is improved, but processing overhead increases
Solution Approach 1:
The patent segments the intra prediction process into distinct paths based on the reference line configuration. When a single reference line is used, RSF is applied; when multiple reference lines are used, the filter is skipped. This segmentation allows the system to handle different prediction scenarios with appropriate processing levels.
Solution Approach 2:
The patent extracts the filter application step from the universal prediction process and makes it conditional. By taking out the RSF operation and applying it only when necessary (single reference line case), the system reduces unnecessary processing overhead while maintaining prediction flexibility through multiple reference line support.
3Measurement precision
If filtering operation is always performed on input samples, then sample accuracy is improved, but hardware resources increase
Solution Approach 1:
The patent applies partial filtering action by performing reference smoothing filtering only on the subset of cases where a single reference line is used (indicated by the index). This partial application of filtering maintains sample accuracy where needed while avoiding redundant filtering operations that would waste hardware resources.
Solution Approach 2:
The system uses the index information already present in the prediction process to self-determine whether filtering is needed. The index serves as a built-in indicator that guides the filter application decision, eliminating the need for additional complex control logic or hardware to determine when filtering should be applied.
Data Source
AI summary
A method of adaptive filtering for multiple reference line of intra prediction in video coding, a video encoding apparatus and video decoding apparatus therewith are provided in the disclosure. In the method of intra prediction in video coding, a method of adaptive filtering is used to dynamically determine operation of filtering is applied to input samples in the intra prediction or not, which can reduce the complexity of the intra prediction in video coding if multiple reference lines are used for the operation of the intra prediction, and also increase the efficiency of compressing performance of ultra-high resolution video.