Image Coding Using SVD Eigenvector Matrices to Reduce Bit Stream Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing SVD decomposition technology in image coding/decoding increases the identifier load of the coded bit stream, reducing coding efficiency due to the need to transmit eigenvector matrices U and V, or their index values, from the encoder to the decoder.
Innovation Solution
Performing singular vector decomposition on a prediction block to obtain eigenvector matrices U and V, and then using these matrices to transform residual data, allowing the encoder to code only the difference information between the eigenvector matrices, thereby reducing the data that needs to be transmitted.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the encoder transmits the matrix U and the matrix V obtained by performing SVD decomposition on the residual data to the decoder, then the decoder can perform inverse transformation to obtain residual data, but the identifier load of the coded bit stream increases and coding efficiency decreases
Solution Approach 1:
The encoder performs SVD decomposition on the prediction block to obtain matrices U and V, and transmits these matrices to the decoder. The decoder uses these transmitted matrices to perform inverse transformation on the residual data. This copying approach allows the decoder to reconstruct the residual data without needing to perform SVD decomposition itself, thereby reducing the identifier load in the coded bit stream while maintaining decoding accuracy.
2Loss of information
If several sets of matrices U and V are determined in the coding/decoding system and the encoder transmits index values corresponding to these matrices, then the data transmission量 is reduced, but the flexibility of the coding/decoding process is limited
Solution Approach 1:
The system determines several sets of matrices U and V with different parameters (e.g., different decomposition conditions or block sizes). The encoder selects the appropriate matrix set based on the characteristics of the current image block and transmits only the index value corresponding to the selected matrices. This allows the system to adapt to different image content while significantly reducing the data transmission volume compared to transmitting the full matrices.
3Productivity
If the encoder performs SVD decomposition on residual data to obtain eigenvector matrices, then transform coding efficiency is improved, but the decoder cannot perform SVD decomposition and needs to receive the matrices from the encoder
Solution Approach 1:
The encoder performs SVD decomposition on the prediction block in advance to obtain the eigenvector matrices U and V. These matrices are then transmitted to the decoder as part of the coding data. This preliminary action allows the decoder to simply use the pre-computed matrices for inverse transformation without needing to implement the complex SVD decomposition algorithm, thereby improving transform coding efficiency while keeping the decoder relatively simple.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present invention discloses an image coding/decoding method, device, and system, where an encoder performs singular vector decomposition on a prediction block corresponding to a to-be-coded image block, to obtain eigenvector matrices U and V of the prediction block; and performs coding processing on residual data according to the eigenvector matrices U and V of the prediction block, where the residual data is a difference between a pixel value of the to-be-coded image block and a pixel value of the corresponding prediction block. This can reduce identifier load of a coded bit stream.