Context-Adaptive Binarization for Video Palette Mode Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current entropy coding methods for video coding, particularly in High Efficiency Video Coding (HEVC) and screen content coding, face inefficiencies in encoding syntax elements for palette mode coded blocks, which affect coding efficiency and bitrate.
Innovation Solution
The method involves generating a binary string for source symbols by binarizing the most significant bit (MSB) index using a unary or truncated unary code and refinement bits using a fixed-length or truncated binary code, with context-adaptive binary arithmetic coding (CABAC) applied to both parts, and selecting contexts based on the palette index to optimize coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional entropy coding methods are used for palette mode coded blocks, then the coding process is simpler, but coding efficiency is reduced and bitrate increases
Solution Approach 1:
The syntax element is divided into two separate binary strings through binarization: a first binary string representing the most significant bits and a second binary string representing the least significant bits. This segmentation allows each part to be coded independently using CABAC, optimizing the coding process for each segment while maintaining overall efficiency.
Solution Approach 2:
Different context models are selectively applied to different parts of the syntax element based on local characteristics. The first binary string uses one context model while the second binary string uses another context model, allowing the coding process to adapt to the specific statistical properties of each segment and improve overall coding efficiency.
2Loss of substance
If standard binarization methods are used, then the implementation is simpler, but bitrate is increased
Solution Approach 1:
The syntax element value is segmented into most significant bits and least significant bits, with each segment converted to a separate binary string. This segmentation enables more efficient entropy coding by allowing context-adaptive binary arithmetic coding to be applied differently to each part, reducing the overall bitrate compared to standard binarization methods.
Data Source
AI summary
A method performs entropy coding and decoding for source symbols generated in a video coding system. The method receives a palette index map for a current block, and determines a number of consecutive pixels having a same palette index as a current palette index in a scanning order through the current block. The method then converts the number of the consecutive pixels minus one to a bin string using a binarization method, and encodes the bin string using context-adaptive binary arithmetic coding (CABAC) by applying a regular CABAC mode to at least one bin of the bin string according to a context adaptively selected depending on the current palette index.