Context Derivation for Video Coding Using Function-Based Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques require large mapping tables to determine contexts for coding bins of last significant coefficients, which can be inefficient in terms of memory usage and processing.
Innovation Solution
Implementing a function-based approach to derive context indexes for coding bins, eliminating the need for mapping tables by using a function of the bin index to determine the context index for CABAC coding of last significant coefficient positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If mapping tables are used to determine contexts for coding bins, then the coding process can be simplified, but memory requirements increase significantly
Solution Approach 1:
The patent extracts the context determination logic from the mapping table and implements it through a function-based approach. Instead of storing pre-computed context indexes in a mapping table, the system computes context indexes dynamically using a function that takes bin index and transform unit size as inputs, thereby removing the memory burden while preserving the coding simplicity.
Solution Approach 2:
The patent replaces the mechanical lookup operation in mapping tables with a computational function. The function `ctx_idx = f(bin_idx, tu_size)` substitutes the table-based mechanism, transforming the context determination from a memory access operation to a mathematical computation that requires minimal memory storage.
2Ease of operation
If mapping tables are used to store context information, then context determination is straightforward, but processing efficiency decreases due to memory access overhead
Solution Approach 1:
The patent substitutes memory access operations with mathematical computations. The context index is calculated using the function `ctx_idx = f(bin_idx, tu_size)` which involves basic arithmetic operations that can be executed faster than memory access, thereby improving processing efficiency while maintaining operational simplicity.
Solution Approach 2:
The system becomes self-sufficient by computing its own context indexes through the function rather than relying on pre-stored values. The function uses readily available parameters (bin index and transform unit size) to generate context indexes on-demand, eliminating the need for external memory resources and reducing processing bottlenecks.
3Adaptability or versatility
If large mapping tables are implemented, then all context scenarios can be covered, but device complexity increases
Solution Approach 1:
The patent implements a universal function `ctx_idx = f(bin_idx, tu_size)` that handles all context determination scenarios regardless of transform unit size or bin index. This single function replaces multiple specialized mapping tables, providing adaptability across different video coding scenarios while reducing device complexity by eliminating the need for multiple data structures.
Solution Approach 2:
Instead of storing data and looking it up (table-based approach), the patent inverts the approach by computing values on-demand through a function. This inversion transforms the system from a data-storage-heavy architecture to a computation-light architecture, reducing device complexity while maintaining comprehensive context coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one example, a device includes a video coder configured to determine a context for entropy coding a bin of a value indicative of a last significant coefficient of a block of video data using a function of an index of the bin, and code the bin using the determined context. The video coder may encode or decode the bin using context-adaptive binary arithmetic coding (CABAC). The function may also depend on a size of the block. In this manner, a table indicating context indexes for the contexts need not be stored by the device.