Adaptive Entropy Coding for Image Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Advanced Display Stream Compression (ADSC) standards face limitations in achieving optimal compression efficiency, particularly in block prediction modes, due to the simplicity of existing entropy coding methods which do not adapt effectively to varying bit-depths and complexities in image data.
Innovation Solution
An adaptive entropy coding method that determines the maximum bit-depth of a group of symbols and selectively uses either Golomb-Rice coding or maximum bit-depth coding based on this threshold, allowing for optimal prefix and suffix bit-length management to minimize coded bit-length and reduce complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple entropy coding methods are used in ADSC, then device complexity is reduced, but compression efficiency deteriorates
Solution Approach 1:
The patent implements dynamic entropy coding by selecting between Golomb-Rice coding and fixed-bit-depth coding based on the maximum bit-depth of symbol groups. This dynamic adaptation allows the system to optimize compression efficiency for different data characteristics while maintaining manageable device complexity through algorithm selection rather than implementing all possible coding methods simultaneously.
Solution Approach 2:
The patent changes the coding parameter (coding method) based on the bit-depth parameter of the data. When the maximum bit-depth of a symbol group exceeds a threshold, Golomb-Rice coding is used; otherwise, fixed-bit-depth coding is applied. This parameter-based adaptation resolves the contradiction by matching coding complexity to data requirements.
2Productivity
If adaptive entropy coding with multiple coding types is implemented, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The patent segments the entropy coding process into distinct coding methods (Golomb-Rice coding and fixed-bit-depth coding) that are selected based on data characteristics. This segmentation allows each coding method to be optimized independently and selected based on suitability, improving compression efficiency without requiring the system to simultaneously handle all coding complexities.
Solution Approach 2:
The patent introduces a selection mechanism that acts as an intermediary between the data and the coding methods. This intermediary evaluates the maximum bit-depth of symbol groups and directs the appropriate coding method, thereby managing the complexity of having multiple coding types while achieving improved compression efficiency.
3Device complexity
If fixed bit-depth coding is used for all symbols, then device complexity is reduced, but loss of information increases due to insufficient precision for high bit-depth symbols
Solution Approach 1:
The patent applies different coding methods to different symbol groups based on their local characteristics (bit-depth). High bit-depth symbol groups receive Golomb-Rice coding with appropriate precision, while low bit-depth symbol groups use simpler fixed-bit-depth coding. This local adaptation ensures information preservation where needed while maintaining simplicity where sufficient.
Data Source
AI summary
A method and apparatus of image coding including adaptive entropy coding are disclosed. According to this method, input pixels associated with a group of symbols generated from image or video data are received. Maximum bit-depth of the group of symbols is then determined. If the maximum bit-depth of the group of symbols is smaller than a first bit-depth threshold, the group of symbols is encoded or decoded using Golomb-Rice coding. If the maximum bit-depth of the group of symbols is greater than or equal to the first bit-depth threshold, the group of symbols is encoded or decoded using second entropy coding, where the second entropy coding is different from the Golomb-Rice coding. Outputs corresponding to encoded or decoded output associated with the group of symbols are provided. The maximum bit-depth of the group of symbols is signaled at the encoder or recovered at the decoder by parsing the bitstream.