Image Encoding Block Transposition for Rate-Distortion Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding techniques, such as those in the HEVC standard, do not optimize the rate-distortion cost effectively, leading to suboptimal compression performance and increased complexity in encoding and decoding processes.
Innovation Solution
A method that predicts a current block using a predictor block, determines a residual data block, and selects a pair of mathematical operations including a permutation and a transformation operation to optimize encoding, specifically using discrete sine transforms (DST) to improve compression performance and reduce calculation resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional block transformations (DCT, DST) are used for encoding, then the transformation can be implemented with fast algorithms and limited algorithmic complexity, but the rate-distortion cost is not optimized effectively
Solution Approach 1:
The patent applies dynamics by making the transformation operation selectable rather than fixed. The encoder can dynamically choose between different transformation operations (first type or second type) based on the characteristics of the residual block, allowing the system to adapt to different content scenarios and optimize rate-distortion cost while maintaining reasonable complexity through predefined transformation options.
Solution Approach 2:
The patent changes the parameter of transformation type selection by introducing a criterion based on residual block characteristics. The encoder evaluates the residual block and selects appropriate transformation parameters (first type or second type) to optimize encoding efficiency, thereby improving rate-distortion cost without excessively increasing algorithmic complexity.
2Productivity
If multiple transformation operations are selected based on optimization criteria, then compression performance is improved, but the complexity of selecting and managing multiple operations increases
Solution Approach 1:
The patent segments the transformation operations into distinct types (first type and second type), each suited for different residual block characteristics. This segmentation allows the encoder to apply the appropriate transformation type based on the specific block being encoded, improving compression performance while keeping the selection process manageable through clear categorization.
Solution Approach 2:
The patent implements dynamic selection between transformation types based on residual block characteristics. The encoder evaluates each block and dynamically chooses the most suitable transformation operation, enabling adaptive optimization of compression performance without requiring complex manual configuration or management of multiple operations.
3Use of energy by moving object
If discrete sine transforms (DST) are used to optimize encoding, then calculation resources are reduced and compression performance is enhanced, but additional signaling information is required
Solution Approach 1:
The patent extracts only the essential transformation type information that needs to be signaled to the decoder. By categorizing transformations into types and signaling only the relevant type indicator, the patent reduces the amount of signaling information required compared to signaling detailed parameters of each possible transformation operation, thus minimizing the cost of signaling while still enabling DST usage for calculation efficiency.
Data Source
AI summary
A method for decoding a data signal representative of at least one image cut into blocks. The method includes, for a current block to be decoded: transforming a residual data block obtained from the data signal; determining a permutation-transformation pair including a transformation mathematical operation and a permutation mathematical operation, the transformation mathematical operation and the permutation mathematical operation belonging to respectively a plurality of permutation operations and a plurality of transformation operations; applying the determined transformation mathematical operation to the data of the residual data block to produce transformed data; applying the determined permutation mathematical operation to the transformed data to produce permuted data, the determined permutation mathematical operation being an operation of transposing the transformed data; and using the permuted data to reconstruct the current block by a predictive decoding.