Compound Orthonormal Transform for Video Coding Complexity Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies, such as HEVC, face inefficiencies in implementing transforms like DST-7 due to higher complexity and implementation costs, particularly with Adaptive Multiple Transforms (AMT), which lack the feature of reusing smaller transforms for larger ones, leading to increased computation burdens and memory requirements.
Innovation Solution
The method involves generating a new multi-transform data structure by embedding non-recursive transforms into a larger recursive transform, specifically replacing the first half of the odd bases of a 2N-point DCT-2 with N-point scaled DST-7 or DCT-8 bases and filling the second half with flipped DST-7 or DCT-8 bases plus inverse signs, creating a Compound Orthonormal Transform (COT) that reduces complexity and shares logic between DCT-2 and DCT-8/DST-7 transforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If Adaptive Multiple Transforms (AMT) with multiple transform types are implemented, then transform flexibility and coding efficiency are improved, but device complexity and implementation cost increase
Solution Approach 1:
The patent merges DCT-2 and DST-7 transforms into a unified transform structure where DST-7 is embedded within the DCT-2 framework. This combining approach allows the system to support multiple transform types (improving adaptability) while using a single integrated implementation structure (reducing device complexity and implementation cost).
Solution Approach 2:
The patent creates a universal transform structure that can perform both DCT-2 and DST-7 transforms through a single implementation. The unified transform uses shared components and logic that can be configured to execute different transform types, making the system multi-functional and reducing the need for separate dedicated hardware for each transform type.
2Productivity
If multiple transform types are supported in AMT, then coding efficiency is improved, but on-chip memory requirements increase
Solution Approach 1:
The patent combines the memory structures for DCT-2 and DST-7 transforms into a unified memory system. By merging the transform kernels and related data structures, the patent reduces total memory requirements while still supporting multiple transform types needed for efficient coding.
Solution Approach 2:
The patent implements a universal memory structure that serves both DCT-2 and DST-7 transforms. This multi-functional memory system can store and retrieve data for different transform types using shared memory resources, reducing the total quantity of memory required compared to having separate dedicated memory for each transform type.
3Measurement precision
If separate transform implementations are used for DCT-2 and DST-7, then transform accuracy is maintained, but computation burden increases
Solution Approach 1:
The patent merges the computation logic for DCT-2 and DST-7 transforms into a single unified implementation. By combining the transform algorithms and sharing computational components, the patent reduces the total computation burden while maintaining the accuracy of both transform types through proper mathematical formulation.
Solution Approach 2:
The patent creates a universal transform implementation that can execute both DCT-2 and DST-7 operations through a single computational framework. This multi-functional approach allows the system to maintain transform accuracy for both types while reducing redundant computations and overall computational power requirements.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method of decoding or encoding that includes generating a new multi-transform data structure by embedding one or more non-recursive transforms into a larger recursive transform. The method may further include receiving information regarding a target data block and determining whether to use a recursive transform or a non-recursive transform. When the determination is to use the recursive transform, the method may include generating the recursive transform using a multi-transform data structure and causing the target data block to be encoded or decoded using the generated recursive transform. If not, the method may include causing the target data block to be encoded or decoded using one of the one or more the non-recursive transforms embedded in the multi-transform data structure.