Context Selection Table Scaling for Video Coding Memory Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video coding standards, such as HEVC, require multiple context selection tables for different TU sizes and color components, leading to increased system cost and complexity due to the need for extensive memory and logic circuits for context selection tables.
Innovation Solution
The method involves generating derived context selection tables for 8×8 TUs from existing tables for 4×4 TUs through down-scaling or up-scaling of indices, allowing sharing of context selection tables among different TU sizes and color components, thereby reducing the number of tables required for hardware implementation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple context selection tables are used for different TU sizes and color components, then the significance map processing accuracy is improved, but the system cost and device complexity increase due to extensive memory and logic circuits required
Solution Approach 1:
The patent applies universality by enabling a single context selection table to serve multiple TU sizes (4×4 and 8×8) and multiple color components (luma and chroma). The context selection table is designed with a unified structure where the context index is derived from coefficient position coordinates that can accommodate different TU dimensions, allowing one table to replace what would traditionally require multiple separate tables.
Solution Approach 2:
The patent changes the parameter representation by using normalized or relative coordinate systems instead of absolute position indices. By expressing coefficient positions as relative coordinates within the TU boundary or using scaled coordinate systems, the same context selection table can map positions for both 4×4 and 8×8 TUs without requiring size-specific tables.
2Manufacturing precision
If separate context selection tables are maintained for different TU sizes, then the coding precision is maintained, but the memory usage and logic circuit requirements increase
Solution Approach 1:
The patent merges multiple separate context selection tables into a single unified table. By combining the context selection logic for 4×4 TUs, 8×8 TUs, luma components, and chroma components into one table structure, the patent reduces the total number of tables from potentially four separate tables to a single shared table, thereby reducing memory usage and simplifying the associated logic circuits.
Solution Approach 2:
The unified context selection table is designed to be multi-functional, serving different TU sizes and color components through a single structure. This universality eliminates the need for separate dedicated tables for each combination of TU size and color component, directly reducing the quantity of memory resources and logic circuits required.
Data Source
AI summary
A method and apparatus for significance map coding for 4×4 TUs (transform units) and 8×8 TUs of video data are disclosed. The method comprises receiving at least one context selection table for a first TU associated with a first color component and generating a derived context selection table for a second TU associated with a second color component from said at least one context selection table. The first TU has a first size comprising of 4×4 and 8×8. The second TU has a second size comprising of 4×4 and 8×8. The second size is different from the first size, the second color component is different from the first color component, or both the second size is different from the first size and the second color component is different from the first color component. The derived context selection table is then used for significance map processing of the second TU.