Video Coding Transform Selection for Non-Zero Reference Lines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face inefficiencies in reducing redundancy in video signals, particularly in intra prediction methods, where less likely prediction directions require more bits, leading to suboptimal compression ratios.
Innovation Solution
The implementation of a transform scheme that determines the primary transform type based on whether a block is predicted using a neighboring reconstructed sample from a non-zero reference line, allowing for the use of DCT-2 or skip modes, and optionally applying DST-7 or DCT-2 transforms depending on the reference line, to optimize transform coefficients and reduce bit requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If traditional intra prediction is used for all blocks, then device complexity is reduced, but compression ratio deteriorates for blocks with less likely prediction directions
Solution Approach 1:
The patent applies parameter changes by dynamically selecting between different transform types (DST-7, DCT-2, or skip mode) based on the intra prediction mode and reference line characteristics. This allows the system to adapt the transform parameters to match the specific block characteristics, improving compression efficiency without significantly increasing overall system complexity.
Solution Approach 2:
The patent implements local quality by applying different transform types to different blocks based on their specific characteristics (intra prediction mode, reference line index). Instead of using a uniform transform approach, the system selectively applies DST-7 for certain prediction modes and reference lines, DCT-2 for others, and skip mode when appropriate, thereby optimizing compression locally for each block type.
2Quantity of substance
If transform skip mode is used for all blocks, then bit usage is reduced, but transform efficiency deteriorates for blocks requiring directional transforms
Solution Approach 1:
The patent dynamically changes the transform parameter (selecting between skip mode, DST-7, and DCT-2) based on the intra prediction mode and reference line characteristics. This allows the system to use skip mode when appropriate (reducing bits) while switching to directional transforms like DST-7 when the block characteristics require them, thus balancing bit usage and transform efficiency.
Solution Approach 2:
The patent introduces dynamics by making the transform type selection adaptive and conditional rather than static. The transform type is determined dynamically based on runtime conditions (intra prediction mode, reference line index), allowing the system to respond to varying block characteristics and optimize the balance between bit usage and transform efficiency for each specific case.
3Device complexity
If a single transform type is used for all intra predicted blocks, then device complexity is reduced, but compression efficiency deteriorates when adapting to different prediction directions
Solution Approach 1:
The patent applies parameter changes by selecting different transform types (DST-7, DCT-2, or skip) based on the intra prediction mode and reference line characteristics. This allows the transform parameters to be adapted to match the specific prediction direction and block characteristics, improving compression efficiency while maintaining manageable device complexity through a limited set of transform options.
Solution Approach 2:
The patent implements local quality by applying different transform types to different blocks based on their specific intra prediction characteristics. Instead of using a uniform transform approach, the system selectively applies transforms that are locally optimized for each block's prediction direction and reference line, thereby improving compression efficiency without requiring complex global transform selection.
4Device complexity
If transform coefficients are not optimized for specific reference lines, then device complexity is reduced, but bit usage increases for blocks using non-zero reference lines
Solution Approach 1:
The patent applies parameter changes by selecting different transform types based on the reference line index. For blocks using non-zero reference lines, the system selects transforms (such as DST-7 for certain modes) that are better suited to the characteristics of those reference lines, thereby optimizing transform coefficients and reducing bit usage without requiring complex adaptive mechanisms.
Data Source
AI summary
According to an aspect of the disclosure, a method of video decoding is performed in a video decoder. In the method, coded information of a current block is received from a coded video bitstream. The coded information includes prediction information of the current block. A determination is made whether the current block is predicted using a neighboring reconstructed sample from a non-zero reference line based on the coded information. A primary transform type is determined based on the determination that the current block is predicted using the neighboring reconstructed sample from the non-zero reference line. A primary transform is performed for a transform block that is partitioned from the current block in accordance with the determined primary transform type.