Integer Sequence Encoding Using Position-Based Difference Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing encoding and decoding methods for sequences of integer values require significant bitstream bandwidth due to lengthy codes, and are inefficient when a large number of different values occur in the sequence.
Innovation Solution
An encoding and decoding apparatus/method that separates a sequence of integer values into two portions, encoding a predetermined value and then encoding differences between neighboring values, using exponential Golomb codes and context-adaptive binary arithmetic coding to optimize bitstream efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional encoding methods are used for sequences of integer values, then the encoding process is simple, but the bitstream bandwidth is large due to lengthy codes
Solution Approach 1:
The sequence of integer values is divided into two portions: a first portion containing integer values up to a predetermined position is encoded using a first exponential Golomb code, and a second portion containing integer values after the predetermined position is encoded using a second exponential Golomb code. This segmentation allows different coding strategies to be applied to different parts of the data, reducing overall bitstream bandwidth while maintaining encoding efficiency.
2Productivity
If conventional encoding methods are used, then all different values are encoded individually, but coding efficiency decreases when a large number of different values occur in the sequence
Solution Approach 1:
The encoding approach changes based on the parameter of position within the sequence. Integer values at different positions are encoded using different exponential Golomb codes with different parameters. Specifically, values at or before the predetermined position use one code configuration, while values after that position use another configuration, optimizing compression for the specific data distribution pattern where values tend to be smaller earlier in the sequence.
Data Source
AI summary
The invention describes an encoding scheme for encoding a sequence of integer values, particularly by first using encoding a value at a predetermined position and then consecutively calculating and encoding differences between neighboring values. A respective decoding scheme is also provided. These encoding and decoding schemes provide a more efficient encoding concept.