Non-Zero Reference Line Intra Prediction for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently reducing redundancy in video signals, particularly in intra prediction modes, where less likely directions require more bits, leading to suboptimal compression ratios and increased bandwidth/storage requirements.
Innovation Solution
The method involves enabling non-angular intra prediction modes for blocks encoded with a non-zero reference line index, allowing for decoding using at least one reference sample from the non-zero reference line, which optimizes bit allocation and compression efficiency by differentiating between zero and non-zero reference lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra prediction uses more bits to represent less likely prediction directions, then prediction accuracy is improved, but bandwidth and storage requirements increase
Solution Approach 1:
The patent segments the reference lines into multiple types (first reference line, second reference line, third reference line) based on their proximity to the current block. Different segmentation strategies are applied: the first reference line uses a default mode, while the second and third reference lines use angular modes with specific direction calculations. This segmentation allows efficient bit allocation by treating different reference lines differently, reducing the total bits needed while maintaining prediction accuracy.
Solution Approach 2:
The patent applies different prediction modes and bit allocation strategies to different local regions (reference lines) based on their specific characteristics. The first reference line (closest to current block) receives default mode treatment, while farther reference lines (second and third) receive angular mode treatment with direction-specific bit allocation. This local quality approach optimizes the balance between prediction accuracy and bit usage for each reference line region.
2Quantity of substance
If compression ratio is increased through lossy compression, then bandwidth and storage requirements are reduced, but video quality deteriorates
Solution Approach 1:
The patent changes the parameter of reference line selection by introducing a reference line index that identifies which reference line to use for prediction. By optimizing the selection of reference lines (choosing from first, second, or third reference lines based on angular mode and direction), the patent improves prediction accuracy without increasing bit rate, thereby maintaining video quality while achieving compression.
3Adaptability or versatility
If multiple reference lines are used for intra prediction, then prediction flexibility is improved, but decoding complexity increases
Solution Approach 1:
The patent introduces dynamic reference line selection through the reference line index, which is determined based on the angular mode and prediction direction. The decoder dynamically selects which reference line (first, second, or third) to use for prediction, allowing flexibility while maintaining manageable complexity through rule-based selection criteria rather than exhaustive search.
Solution Approach 2:
The patent uses a simple reference line index mechanism that can be quickly decoded without complex processing. The reference line index serves as a lightweight indicator that directs the decoder to the appropriate reference line, avoiding the need for complex calculations or multiple decoding paths, thereby reducing decoding complexity while maintaining prediction flexibility.
Data Source
AI summary
A method of video decoding includes retrieving, from a coded video bitstream for a current block encoded in an intra prediction mode, a reference line index indicating a reference line from a plurality of reference lines that include a zero reference line. The method includes determining whether the current block is encoded in a non-angular intra prediction mode. The method includes, in response to the block being encoded in the non-angular intra prediction mode, determining whether the reference line index indicates a non-zero reference line from the plurality of reference lines. The method includes in response to determining that the reference line index is the non-zero reference line, (i) enabling non-angular intra prediction modes for the non-zero reference line, and (ii) decoding the current block with at least one reference sample from the non-zero reference line.