Selective Transform Omission in Video Decoding CTUs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video coding schemes, such as H.264/MPEG-4 and HEVC, face inefficiencies in restoring predictive residues for Transform Units (TUs), which affects the overall decoding and encoding process.
Innovation Solution
An image decoding and encoding device that splits pictures into Coding Tree Units (CTUs) and applies multiple transforms, including a first and second transform, to sub-blocks within the CTU, with the option to omit transforms based on flags or CTU size, allowing for efficient restoration of predictive residues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single transform is applied to the entire Coding Tree Unit, then the device complexity is reduced, but the manufacturing precision of predictive residue restoration deteriorates
Solution Approach 1:
The Coding Tree Unit is divided into multiple sub-blocks, and different transform processing can be applied to different sub-blocks based on their characteristics. This segmentation allows the system to achieve high restoration precision for different regions while managing overall processing complexity through selective application of transform operations.
Solution Approach 2:
Different transform types or parameters are applied to different sub-blocks within the CTU based on local characteristics such as prediction mode, block size, or content properties. This local quality approach enables optimized restoration precision for each sub-region without uniformly increasing complexity across the entire CTU.
2Manufacturing precision
If multiple transforms are applied to sub-blocks within the Coding Tree Unit, then the manufacturing precision of predictive residue restoration is improved, but the device complexity increases
Solution Approach 1:
Instead of applying multiple transforms to all sub-blocks uniformly, the system selectively applies additional transform processing only to specific sub-blocks where it provides the most benefit. This partial action approach improves restoration precision for critical regions while limiting the increase in overall device complexity.
Solution Approach 2:
The transform processing configuration is made dynamic and adaptive, allowing the system to adjust the number and type of transforms applied to each sub-block based on local characteristics. This dynamic approach enables the system to optimize between precision and complexity on a per-sub-block basis rather than using a fixed complex configuration for the entire CTU.
3Manufacturing precision
If transform processing is applied to all sub-blocks, then the manufacturing precision is improved, but the loss of time increases
Solution Approach 1:
The system applies detailed transform processing only to sub-blocks where it is most needed based on local characteristics, rather than uniformly processing all sub-blocks. This local quality approach maintains high restoration precision for critical regions while reducing the overall time loss by skipping or simplifying processing in regions where it provides minimal benefit.
Solution Approach 2:
Instead of applying full transform processing to all sub-blocks, the system applies transform processing partially to only those sub-blocks where it provides significant improvement. This selective partial action reduces the total processing time while maintaining adequate restoration precision for the most important regions.
4Productivity
If transform processing is omitted based on flags or CTU size, then the productivity is improved, but the manufacturing precision deteriorates
Solution Approach 1:
The system uses flags and CTU size criteria to identify specific sub-blocks or regions where transform processing can be omitted without significantly impacting overall restoration precision. This local quality approach allows productivity improvements through selective omission while maintaining adequate precision in regions where transform processing is most critical.
Solution Approach 2:
Instead of completely omitting transform processing, the system applies it partially to only those sub-blocks where it provides the most benefit, as determined by flags and CTU size criteria. This partial application approach improves productivity by skipping processing in less critical regions while maintaining adequate restoration precision where needed.
Data Source
AI summary
The present invention avoids waste caused by performing both a Secondary Transform and an Adaptive Multiple Core Transform. Provided is a device including: a core transform unit (1521) that can perform an Adaptive Multiple Core Transform on a Coding Tree Unit; and a Secondary Transform unit (1522) that can perform, before the Adaptive Multiple Core Transform, a Secondary Transform on at least any one of sub-blocks included in the Coding Tree Unit. The device omits any of the Adaptive Multiple Core Transform and the Secondary Transform in accordance with at least any of a flag associated with the Adaptive Multiple Core Transform and a flag associated with the Secondary Transform, or in accordance with a size of the Coding Tree Unit.