Binary Arithmetic Coding for Low-Complexity Integer Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional binary arithmetic coding schemes face challenges in efficiently encoding and decoding non-negative integers with low computational complexity, particularly in handling multiple MPS's within a single renormalization cycle and preserving the ability to encode conventional bins.
Innovation Solution
The proposed solution extends table-based binary arithmetic encoding and decoding by using a unary binarization of non-negative integer values, processing multiple MPS's within a single renormalization cycle, and configuring code lengths similar to Exponential-Golomb or Golomb-Rice codes, with specific steps for encoding and decoding that involve determining maximum numbers of MPS's and LPS's, renormalization, and updating interval widths and offsets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional binary arithmetic coding schemes are used, then the ability to encode conventional bins is preserved, but the computational complexity increases when handling multiple MPS's within a single renormalization cycle
Solution Approach 1:
The patent segments the encoding process into distinct phases: determining the maximum number of MPS's that can be encoded before renormalization is needed, encoding those MPS's in batch, and then performing renormalization. This segmentation allows the decoder to process multiple MPS's efficiently within a single renormalization cycle while maintaining compatibility with conventional bin encoding.
Solution Approach 2:
The patent performs preliminary determination of the maximum number of MPS's (cmps) that can be encoded before renormalization is required. By calculating cmps = (R - Rmin) / 2^s in advance, the system prepares the encoding state beforehand, allowing efficient batch processing of multiple MPS's without increasing overall computational complexity.
2Productivity
If multiple MPS's are processed within a single renormalization cycle, then encoding efficiency improves, but the difficulty of detecting and measuring the maximum number of MPS's increases
Solution Approach 1:
The patent uses parameter changes to simplify detection: it introduces the parameter s (where 2^s represents the partial interval width for LPS) and uses it to calculate cmps = (R - Rmin) / 2^s. This parameter-based approach transforms a complex detection problem into a simple arithmetic operation, making it easy to determine how many MPS's can be encoded before renormalization is needed.
Solution Approach 2:
The patent introduces an intermediary variable cmps (maximum number of MPS's before renormalization) that mediates between the current interval width R and the minimum allowed width Rmin. This intermediary simplifies the detection process by providing a clear numerical threshold that indicates when renormalization must occur.
3Loss of energy
If code lengths are configured similar to Exponential-Golomb or Golomb-Rice codes, then compression efficiency improves, but the adaptability to encode both conventional bins and non-negative integers decreases
Solution Approach 1:
The patent creates a universal arithmetic coding scheme that can handle both conventional bins and non-negative integers. By using unary binarization for non-negative integers (mapping value c to c MPS's followed by an LPS) and maintaining compatibility with conventional bin encoding, the system achieves multi-functionality without sacrificing compression efficiency or requiring separate coding paths.
Data Source
AI summary
A binary arithmetic coding scheme is extended by a functionality to encode and decode non-negative integer values with particular low computational complexity.