Intra Base Layer Transform Selection in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding techniques do not allow for transform selection based on the characteristics of the data being transformed, particularly in scalable video coding (SVC) when using inter-layer prediction modes, which limits coding efficiency and computational resource utilization.
Innovation Solution
The proposed solution involves selecting between different transforms, such as DCT or DST, based on the transform unit size and color component type of the enhancement layer video information, allowing for adaptive transformation during encoding and decoding processes without requiring new hardware or low-level changes to existing HEVC processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a single fixed transform is used for all video blocks, then the implementation is simple, but coding efficiency is limited
Solution Approach 1:
The patent implements dynamic transform selection where the transform type (DST or DCT) is chosen based on the characteristics of each video block, specifically the intra prediction mode and block size. This allows the system to adapt transform selection to local data characteristics rather than using a fixed transform for all blocks, thereby improving coding efficiency while maintaining reasonable implementation complexity through clear selection criteria.
Solution Approach 2:
The patent applies different transform types to different regions of the video based on local characteristics. Specifically, DST is applied to certain intra-predicted blocks while DCT is applied to others, depending on the prediction mode and block size. This localized approach ensures that each region is transformed using the most appropriate method for its specific characteristics, improving overall coding efficiency.
2Productivity
If transform selection is implemented based on data characteristics, then coding efficiency is improved, but computational resources increase
Solution Approach 1:
The patent changes the parameter being transformed based on specific conditions (intra prediction mode and block size). By selecting between DST and DCT transforms according to these parameters, the system optimizes coding efficiency for different block types while avoiding the computational overhead of evaluating multiple transform options for every block. The selection criteria are designed to be computationally efficient.
Solution Approach 2:
The patent applies computational resources selectively by using different transforms only where beneficial. Instead of implementing a complex adaptive system for all blocks, the patent applies DST specifically to certain intra-predicted blocks where it provides improvement, while using the simpler DCT for other blocks, thus optimizing the balance between coding efficiency and computational cost.
3Adaptability or versatility
If existing HEVC processes are modified to support transform selection, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent introduces dynamic transform selection into the existing HEVC framework by adding conditional logic based on intra prediction mode and block size. The selection process follows clear rules that determine whether to use DST or DCT, allowing the system to adapt to different video content characteristics while integrating smoothly with the existing HEVC encoding and decoding processes without requiring fundamental architectural changes.
Data Source
AI summary
An apparatus for coding video information according to certain aspects includes a memory unit and a processor in communication with the memory unit. The memory unit is configured to store video information associated with a base layer and an enhancement layer. The processor is configured to, in response to determining that the video information associated with the enhancement layer is to be determined based upon the video information associated with the base layer, select between a first transform and a second transform based at least in part on at least one of a transform unit (TU) size and a color component type of the enhancement layer video information.