Recursive Intra Region Entropy Coding With Contextual Flag Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods face inefficiencies in entropy coding due to the lack of utilizing coded information as context for probability models, leading to increased overhead in signaling flags for video characteristics.
Innovation Solution
Use coded information such as previous instances of signaled flags, block sizes, and neighboring block information as context for entropy encoding, to improve the accuracy and efficiency of entropy coding by reducing the overhead required for signaling flags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional entropy coding without context is used for signaled flags, then the coding process is simple, but the overhead for signaling flags increases and compression efficiency decreases
Solution Approach 1:
The patent applies feedback by using previously decoded signaled flags and block size information as context for entropy coding current flags. This feedback mechanism allows the encoder to leverage historical data and spatial relationships to improve coding efficiency, reducing the overhead required for signaling while maintaining or increasing compression performance.
Solution Approach 2:
The patent changes the parameters of the entropy coding process by introducing context-dependent probability models that vary based on block size, neighboring block characteristics, and previous flag values. This parameter adaptation enables more efficient coding by adjusting the probability distribution according to the specific coding situation, thereby reducing signaling overhead without significantly increasing complexity.
2Productivity
If coded information is used as context for entropy encoding, then the accuracy and efficiency of entropy coding improve, but the complexity of the coding process increases
Solution Approach 1:
The patent segments the entropy coding process into multiple context models, each handling different coding situations (e.g., different block sizes, different neighboring block configurations). This segmentation allows the system to manage complexity by dividing the overall coding task into manageable sub-tasks, each with its own specialized context handling, while achieving high compression efficiency through optimized per-context coding.
Solution Approach 2:
The patent performs preliminary actions by pre-defining context models and probability distributions based on statistical analysis of block sizes and neighboring block characteristics. This preliminary preparation allows the actual coding process to proceed more efficiently by relying on pre-computed context information, reducing the computational burden during real-time encoding while maintaining high accuracy.
Data Source
AI summary
An example method of video coding includes receiving a video bitstream including a plurality of coding blocks; identifying a coding region that includes two or more coding blocks of the plurality of coding blocks that are encoded in a first prediction mode. The method also includes entropy decoding a signaled flag indicating a prediction mode for the coding region, the entropy decoding using coded information including one or more of: previous instances of the signaled flag, a block size of a current coding block, a block size group of the current coding block, respective block sizes of a set of neighboring coding blocks, respective block size groups of the set of neighboring coding blocks, and signaled flags for the set of neighboring coding blocks; and reconstructing the two or more coding blocks according to a value of the signaled flag for the coding region.