Video Decoder Context Sharing for Lower Storage Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as HEVC, struggle to further compress videos without sacrificing picture quality, leading to a significant burden on communications networks and data storage.
Innovation Solution
A video decoding method that shares a context model for syntax elements in a current block, reducing the need for multiple context models and minimizing storage space in decoders by using a single context model for syntax elements 1 and 2, and employing context index determination based on neighboring block values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple context models are used for different syntax elements in a current block, then decoding accuracy is improved, but storage space required by the decoder increases
Solution Approach 1:
The patent merges multiple context models into a single shared context model that is common to all syntax elements in a current block. This is achieved by using a unified context index determination method that selects from a limited set of preset context models (e.g., first, second, and third context models) based on syntax element types and neighboring block information, rather than maintaining separate context models for each syntax element.
Solution Approach 2:
The single shared context model serves multiple functions by being applicable to different syntax elements (such as motion vector prediction flags, merge mode indicators, and other block-level syntax elements) through a universal context index determination mechanism that adapts to different syntax element types using neighboring block context.
2Measurement precision
If multiple context models are maintained for different syntax elements, then decoding precision is improved, but device complexity increases
Solution Approach 1:
The patent combines multiple context model management functions into a single simplified context model structure. The context index determination unit consolidates the logic for selecting appropriate context models by using a standardized set of preset context models and a unified selection criterion based on syntax element types and neighboring block characteristics, reducing the complexity of context model management.
Solution Approach 2:
The patent extracts the context model selection logic from the syntax element decoding process itself. By determining the context index before decoding based on syntax element type and neighboring block information, the complex context model selection is separated from the decoding operation, simplifying the overall decoder architecture.
3Quantity of substance
If a single context model is used for all syntax elements, then storage space is reduced, but decoding efficiency decreases
Solution Approach 1:
The patent applies local quality by differentiating context model selection based on syntax element types and spatial location. Different preset context models are selected according to the specific syntax element being decoded (e.g., motion vector-related vs. prediction mode-related elements) and the characteristics of neighboring blocks, allowing optimized decoding for each local context while using a single shared model structure.
Solution Approach 2:
The patent introduces dynamics by making the context model selection adaptive rather than static. The context index is dynamically determined based on the current syntax element type and neighboring block characteristics, allowing the single context model to adapt its behavior to different decoding scenarios, thereby maintaining high decoding efficiency despite using a unified model structure.
Data Source
AI summary
A video decoding method includes obtaining a to-be-entropy-decoded syntax element in a current block by parsing a received bitstream, where the to-be-entropy-decoded syntax element includes a syntax element 1 or a syntax element 2 in the current block, obtaining a context model corresponding to the to-be-entropy-decoded syntax element, where both of a context model corresponding to the syntax element 1 and a context model corresponding to the syntax element 2 are determined from the same preset context model set, entropy decoding the to-be-entropy-decoded syntax element based on the context model corresponding to the to-be-entropy-decoded syntax element, and obtaining a reconstructed image of the current block based on the syntax element obtained by entropy decoding.