Quantization Scaling Matrix Update for Image Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image and video coding schemes require a large number of bits to transmit quantization scaling matrix values, which reduces coding efficiency.
Innovation Solution
An image coding method and decoding method that compute a new quantization scaling matrix based on an old matrix and an update parameter, allowing for reduced information transmission by using a prediction matrix and delta matrix, thereby decreasing the amount of data needed in the coded stream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If quantization scaling matrix values are transmitted in a picture header using conventional methods, then the decoder can obtain accurate quantization parameters, but the number of bits required increases significantly, reducing coding efficiency
Solution Approach 1:
The patent applies preliminary action by transmitting only the difference values between consecutive quantization scaling matrices rather than the full matrix values. The decoder reconstructs the complete matrices by accumulating these differences sequentially, thereby reducing the bit rate while maintaining accuracy.
Solution Approach 2:
The patent changes the parameter representation from absolute quantization scaling matrix values to relative difference values. This transformation reduces the magnitude and variability of the transmitted parameters, enabling more efficient entropy coding and reducing the overall bit rate requirement.
2Reliability
If all quantization scaling matrix values are coded in the picture header, then the decoder has complete information for accurate decoding, but the device complexity and processing overhead increase
Solution Approach 1:
The decoder performs preliminary accumulation of difference values to reconstruct the full quantization scaling matrices. This approach maintains decoding accuracy while simplifying the header processing by dealing with smaller difference values rather than complete matrices.
3Productivity
If the quantization scaling matrix size is increased to improve compression quality, then the compression efficiency improves, but the number of bits required to transmit the matrix values increases
Solution Approach 1:
By transmitting difference values from the previous matrix, the patent enables the use of larger quantization scaling matrices without proportionally increasing the bit rate. The difference values for larger matrices remain relatively small and can be efficiently encoded.
Solution Approach 2:
The patent transforms the transmission parameter from absolute matrix values to relative differences, which reduces the information content that needs to be transmitted regardless of matrix size, thereby enabling larger matrices at the same bit rate.
Data Source
AI summary
An image decoding method including: obtaining an old quantization scaling matrix which is a decoded quantization scaling matrix and is used for decoding a new quantization scaling matrix; obtaining, from the coded stream, an update parameter indicating an amount of change in the new quantization scaling matrix with respect to the old quantization scaling matrix; decoding the new quantization scaling matrix using the old quantization scaling matrix obtained in the obtaining of an old quantization scaling matrix and the update parameter obtained in the obtaining of an update parameter; and decoding the coded image using the new quantization scaling matrix decoded in the decoding of the new quantization scaling matrix.