Video Decoding Transform Type Determination Using Block Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques lack an efficient method to determine the primary transform type for decoding target blocks, especially in scenarios where multiple transform sets are applied, leading to suboptimal compression performance.
Innovation Solution
A video decoding method that obtains parameters indicating the applicability of multiple transform sets (MTS) to decoding target blocks, determines the transform type based on these parameters, sets a zero-out area, and performs an inverse transform accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple transform sets are applied to decoding target blocks, then compression performance can be improved, but the complexity of determining the primary transform type increases
Solution Approach 1:
The patent applies preliminary action by pre-establishing a mapping relationship between transform types and block size characteristics before decoding. The encoder determines and signals the primary transform type based on block size criteria (e.g., 4x4, 8x8, 16x16, 32x32 blocks), and the decoder uses this pre-defined mapping to quickly determine the transform type without complex real-time analysis, thus improving compression performance while avoiding determination complexity
Solution Approach 2:
The patent utilizes parameter changes by varying the transform type parameter according to block size parameters. Different transform types (e.g., DCT-II, DST-7, DCT-8) are selected based on the block dimensions, allowing the system to adapt the transform parameter to the specific block size, thereby optimizing compression performance for different block sizes while maintaining a systematic determination approach
2Device complexity
If the transform type is determined based on block size, then the determination process becomes simpler, but the adaptability to different block characteristics may be reduced
Solution Approach 1:
The patent applies local quality by assigning different transform types to different block size regions. Each block size category (4x4, 8x8, 16x16, 32x32) is treated as a local region with specific transform characteristics. The system determines transform types locally based on block size, allowing each block region to receive the most appropriate transform while maintaining a simple overall determination process through hierarchical classification
Data Source
AI summary
Provided are a video decoding method and device. This specification provides a video decoding method comprising the steps of: acquiring a parameter indicating whether a multiple transform set is applicable to a block to be decoded, as well as information about the width of the block to be decoded and the height of the block to be decoded; determining the transform type of the block to be decoded on the basis of at least one of the parameter indicating whether a multiple transform set is applicable, or the information about the width and height of the block to be decoded, and setting a zero-out region of the block to be decoded; and inverse-transforming the block to be decoded on the basis of the zero-out region of the block to be decoded and the result of determining the transform type.