Adaptive Transform Selection in Scalable Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies, such as HEVC SVC extensions, face inefficiencies in residual prediction, particularly in multi-loop decoding, where inter-layer residual prediction methods can be improved for better compression performance.
Innovation Solution
The implementation of a method that determines a transform function based on a threshold value related to the difference between video block information and neighboring video block information, using alternative transforms like discrete-sine transforms (DST) or discrete-cosine transforms (DCT) for Intra BL mode and Generalized Residual Prediction (GRP) blocks, instead of the primary DCT, to enhance coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If primary DCT transform is used for all video blocks, then implementation complexity is low, but coding efficiency is insufficient for blocks with specific characteristics (Intra BL mode and GRP blocks)
Solution Approach 1:
The patent applies different transform functions to different types of video blocks based on their specific characteristics. Intra BL mode blocks and GRP blocks use alternative transforms (ADST, ASDCT, or KLT), while other blocks use the primary DCT transform. This localized adaptation optimizes coding efficiency for specific block types without unnecessarily increasing complexity for all blocks.
Solution Approach 2:
The patent introduces dynamic transform selection based on block type and characteristics. The transform function is not fixed but adapts according to the coding mode (Intra BL, GRP, or other modes) and block properties. This dynamic approach allows the system to choose the most appropriate transform for each block, improving overall coding efficiency.
2Loss of energy
If alternative transforms (ADST, ASDCT, KLT) are used for Intra BL mode and GRP blocks, then energy compaction is improved, but computational complexity increases
Solution Approach 1:
The patent applies computationally more complex alternative transforms (ADST, ASDCT, KLT) only to specific block types (Intra BL mode and GRP blocks) where they provide significant energy compaction benefits. The majority of blocks continue to use the simpler primary DCT transform, thus limiting the increase in overall computational complexity while achieving energy compaction improvements where needed.
Solution Approach 2:
The patent changes the transform function parameter based on block type characteristics. For Intra BL mode and GRP blocks, alternative transforms with different mathematical properties are selected to achieve better energy compaction. This parameter change is conditional and targeted, balancing the trade-off between energy compaction improvement and computational complexity increase.
3Manufacturing precision
If transform function is selected based on block characteristics, then coding precision is improved, but processing complexity increases
Solution Approach 1:
The patent implements local adaptivity by selecting transform functions based on specific block characteristics (coding mode, block type). Intra BL mode blocks and GRP blocks receive specialized transform treatment for improved coding precision, while other blocks use the standard transform. This localized precision improvement avoids the need for complex decision-making in every block processing case.
Solution Approach 2:
The patent performs preliminary classification of blocks into different categories (Intra BL mode, GRP mode, other modes) before applying transforms. This preliminary action allows the system to pre-determine the appropriate transform function based on block characteristics, reducing processing complexity during the actual transform operation while maintaining high coding precision.
Data Source
AI summary
An apparatus for coding video data according to certain aspects includes a memory and a processor in communication with the memory. The memory stores video block information. The video block information includes reference layer block information. The processor determines, based on a parameter of the video block information, a transform function that may be used to code the video block information. The processor may encode or decode the video block information. The transform function may be an alternative transform when the parameter is a predetermined value and a primary transform when the parameter is not the predetermined value. The alternative transform includes one of: a discrete-sine-transform (DST), a Type-I DST, a Type-III DST, a Type-IV DST, a Type-VII DST, a discrete-cosine-transform (DCT), a DCT of different types, and a Karhunen-Loeve transform (KLT).