CABAC Context Inheritance to Reduce Intra Prediction Bit Usage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in representing intra prediction directions, particularly those that are statistically less likely to occur in video content, leading to increased bit usage and reduced compression ratios.
Innovation Solution
Implementing context-based adaptive binary arithmetic coding (CABAC) with inherited context initialization, where context models are initialized based on stored context states from decoded portions of the video sequence, allowing for more efficient encoding and decoding of intra prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional entropy coding is used for intra prediction directions, then all directions are coded with the same number of bits, but this leads to increased bit usage for statistically less likely directions and reduced compression ratios
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the CABAC context model parameters based on the statistical likelihood of different intra prediction directions. Context models are initialized with probability values that reflect the statistical distribution of prediction directions, allowing more bits to be allocated to likely directions and fewer bits to unlikely directions, thereby optimizing the compression ratio while adapting to the actual content characteristics
Solution Approach 2:
The patent implements dynamics by making the entropy coding parameters adaptive rather than static. The context models are dynamically updated during decoding based on previously decoded directions and content characteristics, allowing the coding system to adapt to the statistical properties of the current video content and achieve better compression efficiency for different types of content
2Measurement precision
If context models are reinitialized for each decoded portion, then coding accuracy is maintained, but this increases computational complexity and processing time
Solution Approach 1:
The patent applies preliminary action by pre-initializing context models with statistically derived probability values before decoding begins. This preliminary setup allows the decoder to start with optimized parameters that reflect typical video content characteristics, reducing the need for extensive real-time adaptation while maintaining coding accuracy
Solution Approach 2:
The patent implements self-service by enabling the context models to automatically adapt to the current content characteristics during decoding without requiring complex external control. The models use the decoded content itself to update their probability estimates, allowing the system to self-adjust to maintaining optimal coding accuracy while reducing computational overhead compared to full reinitialization
Data Source
AI summary
Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes receiving circuitry and processing circuitry. The processing circuitry determines, a use of an inherited context initialization in a context-based adaptive binary arithmetic coding (CABAC) for decoding a current portion of a video sequence carried by a bitstream. The processing circuitry decodes, from the bitstream, a syntax value that indicates a decoded portion in the video sequence for a context inheriting, and initializes context models for the CABAC based on stored context states associated with the decoded portion. Further, the processing circuitry decodes, the current portion from the bitstream according to the context models that are initialized based on the stored context states associated with the decoded portion.