Video Encoder Adaptive Transform Basis Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods require significant processing to find the best combination of transform bases and parameters, leading to increased processing loads without optimizing compression efficiency.
Innovation Solution
The proposed encoder and decoder systems selectively perform first and second transforms based on intra prediction modes and transform basis selection, allowing for reduced processing by skipping unnecessary transforms and optimizing transform parameter search.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional video encoding methods perform exhaustive search for best transform basis combination, then compression efficiency is improved, but processing load increases significantly
Solution Approach 1:
The patent implements dynamic transform basis selection where the encoder adaptsively chooses between first transform basis and second transform basis based on intra prediction mode and block size. This dynamic adaptation allows the system to optimize compression efficiency for different coding scenarios while avoiding unnecessary processing by selecting appropriate transforms based on current conditions.
Solution Approach 2:
The patent changes the transform basis parameter based on intra prediction mode and block size conditions. When intra prediction mode is directional and block size is 64x64 or smaller, the system uses first transform basis; otherwise it uses second transform basis. This parameter change strategy optimizes compression efficiency while reducing processing load by avoiding exhaustive search in all cases.
2Manufacturing precision
If the encoder performs both first transform and second transform for all blocks, then compression efficiency is maximized, but processing time increases
Solution Approach 1:
The patent applies partial transform based on specific conditions. Instead of always performing both first transform and second transform, the system performs transforms selectively: first transform is applied when intra prediction mode is directional and block size is 64x64 or smaller; second transform is applied when intra prediction mode is non-directional or block size is larger. This partial action approach reduces processing time while maintaining adequate compression efficiency.
Solution Approach 2:
The encoder performs preliminary determination of intra prediction mode and block size before selecting transform basis. This preliminary action allows the system to pre-determine which transform to apply, avoiding unnecessary transform operations and reducing processing time while maintaining compression efficiency through informed transform selection.
3Manufacturing precision
If the encoder uses adaptive basis selection mode with multiple transform bases, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The patent segments the transform selection process into distinct cases based on intra prediction mode and block size. The encoder divides the coding space into different segments (directional prediction with small blocks vs. non-directional prediction with large blocks) and applies different transform bases to each segment. This segmentation simplifies the overall complexity by providing clear decision rules rather than requiring complex adaptive algorithms.
Data Source
AI summary
When intra prediction is to be used, an encoder determines whether an adaptive basis selection mode is to be used and whether a size of the current block matches a determined size. When the adaptive basis selection mode is to be used and the size matches the determined size, the encoder fixes the first transform basis for the current block to a first determined transform basis in the adaptive basis selection mode, and generates first transform coefficients by performing first transform of residual signals of the current block using the first determined transform basis; quantizes the first transform coefficients when an intra prediction mode for the current block is a determined mode; and generates second transform coefficients by performing second transform of the first transform coefficients using a second transform basis, and quantizes the second transform coefficients, when the intra prediction mode is not the determined mode.