Virtual Reference Line Intra Prediction for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques face inefficiencies in utilizing multiple reference lines for intra prediction, as they often consider only one candidate pixel line, limiting the diversity and accuracy of prediction, especially with increasing image sizes and resolutions.
Innovation Solution
A method and apparatus that generate a virtual reference line by combining multiple pixel rows and columns with high spatial similarity, allowing for the selection of an optimal reference line among them for intra-prediction of current blocks, enhancing video coding efficiency and quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If only one candidate pixel line is considered as a reference line in MRL technique, then the device complexity is reduced, but the manufacturing precision of prediction accuracy deteriorates
Solution Approach 1:
The patent segments the reference line selection process into two distinct parts: (1) generation of a virtual reference line by combining multiple candidate pixel lines, and (2) selection of the optimal reference line from the virtual reference line and original candidate lines. This segmentation allows comprehensive utilization of multiple pixel lines while maintaining manageable processing complexity through structured selection criteria.
Solution Approach 2:
The patent merges multiple candidate pixel lines (first pixel line and second pixel line) to generate a virtual reference line. This combining approach integrates information from multiple sources to create a more accurate reference, directly improving prediction accuracy while the subsequent selection process controls the complexity increase.
2Productivity
If multiple pixel lines are utilized for intra prediction, then the productivity of video coding efficiency is improved, but the device complexity increases
Solution Approach 1:
The patent implements dynamic reference line selection by determining an optimal reference line index based on block characteristics, prediction modes, and gradient calculations. This dynamic approach adapts the reference line selection to specific coding contexts, improving video coding efficiency while avoiding the need to process all possible reference lines uniformly, thus controlling device complexity.
Solution Approach 2:
The patent performs preliminary actions by generating a virtual reference line and calculating gradient values before the actual prediction process. This preliminary preparation organizes the multiple pixel lines into a structured format with pre-computed metrics, enabling efficient subsequent selection and prediction operations, thereby improving overall coding efficiency without proportionally increasing complexity during the main processing stage.
3Measurement precision
If the number of available candidate pixel lines is increased, then the measurement precision of reference pixel accuracy is improved, but the difficulty of detecting and measuring increases
Solution Approach 1:
The patent employs feedback mechanisms by calculating gradient values that measure the accuracy of different reference line candidates and using this information to select the optimal reference line. The gradient calculation provides quantitative feedback on reference pixel accuracy, and this feedback is used to guide the selection process, making the measurement precise while managing the complexity through objective criteria rather than subjective judgment.
Data Source
AI summary
A method and an apparatus are disclosed for video coding using a virtual reference line. In the disclosed embodiments, a video decoding device decodes a virtual reference line usage flag and a reference mode index. When the virtual reference line usage flag is true, the video decoding device generates a reference mode list and derives an intra prediction mode of the current block from the reference mode list by using the reference mode index. In addition, the video decoding device generates the virtual reference line from a plurality of preset reference lines based on the virtual reference line usage flag and generates a prediction block of the current block using the virtual reference line based on the intra prediction mode.