Transform Coefficient Entropy Coding With Dynamic Rice Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution, high-quality images and videos, particularly in virtual reality and augmented reality, necessitates a more efficient compression technique to reduce transmission and storage costs while maintaining image quality.
Innovation Solution
A method and apparatus for enhancing video/image coding efficiency by deriving a Rice parameter based on the maximum number of context-coded bins for transform coefficients, improving residual coding efficiency and reducing memory size without additional look-up tables or signaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If conventional entropy encoding methods are used for high-resolution video, then transmission and storage costs increase, but image quality is maintained
Solution Approach 1:
The patent changes the parameter for determining the number of context models from fixed (based on coefficient position) to dynamic (based on coefficient level magnitude). This allows the encoding system to adapt to the actual distribution characteristics of transform coefficients, improving compression efficiency without sacrificing reconstruction quality
Solution Approach 2:
The patent introduces dynamic adaptation of context model count based on the magnitude of transform coefficient levels. Instead of using a static number of context models for all coefficients, the system dynamically adjusts the number of context models used for encoding each coefficient based on its level value, optimizing the trade-off between compression rate and quality
2Measurement precision
If the number of context models is increased for better encoding precision, then device complexity increases
Solution Approach 1:
The patent changes the parameter for determining context model count from fixed to dynamic, based on coefficient level magnitude. This allows the system to use more context models only when necessary (for larger coefficient values that benefit from finer probability estimation), while using fewer context models for small coefficients, thus balancing precision and complexity
Solution Approach 2:
The patent applies different numbers of context models to different transform coefficients based on their local characteristics (level magnitude). Coefficients with larger magnitudes use more context models for higher precision encoding, while coefficients with small magnitudes use fewer context models, optimizing the overall balance between precision and complexity
3Manufacturing precision
If high-resolution video is transmitted without compression, then transmission and storage costs increase significantly
Solution Approach 1:
The patent applies dynamic parameter adjustment in the entropy encoding stage, changing how context models are selected and applied based on transform coefficient characteristics. This improves compression efficiency by better matching the probability models to the actual data distribution, achieving higher compression ratios while maintaining quality
Solution Approach 2:
The patent uses transform coefficients as an intermediary representation between the original image data and the final encoded bitstream. By operating on transform coefficients with dynamically adjusted context models, the system achieves efficient compression while preserving the ability to reconstruct high-quality images
Data Source
AI summary
A method by which a decoding device decodes a video, according to the present document, can comprise the steps of: acquiring, from a bitstream, information that indicates the level value of a transform coefficient in a current block; deriving, on the basis of whether the number of context coded bins maximally usable for the transform coefficient is greater than or equal to a threshold value, a rice parameter for the information that indicates the level value of the transform coefficient; deriving, on the basis of the rice parameter, a bin string for the information that indicates the level value of the transform coefficient; and deriving the level value of the transform coefficient on the basis of the bin string.