Distributed arithmetic coding with interval swapping

In distributed source coding, ambiguity is usually introduced in the encoding process as it allows multiple plaintext sequences encoded to the same codeword. All these plaintext sequences are decodable and are considered as candidates at the decoder. With the help of side information, the decoder is...

Full description

Saved in:
Bibliographic Details
Published in:Signal processing Vol. 116; pp. 29 - 37
Main Authors: Zhou, Junwei, Wong, Kwok-Wo, Yang, Yanchao
Format: Journal Article
Language:English
Published: Elsevier B.V 01.11.2015
Subjects:
Correlated data
Data compression
Distributed source coding
Slepian-Wolf coding
Correlated data
Slepian-Wolf coding
Data compression
Distributed source coding
ISSN:0165-1684, 1872-7557
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In distributed source coding, ambiguity is usually introduced in the encoding process as it allows multiple plaintext sequences encoded to the same codeword. All these plaintext sequences are decodable and are considered as candidates at the decoder. With the help of side information, the decoder is able to determine which sequence in the candidate set is the best choice. Both the cardinality and the minimum Hamming distance of the candidate set are significant to the decoding performance. In this paper, a Slepian–Wolf code based on arithmetic coding is studied. By employing the interval swapping technique, a linear code is incorporated into binary arithmetic coding. The incorporated linear code improves the minimum Hamming distance within the candidate set which leads to a lower bit error probability. Moreover, binary arithmetic coding exploits the a priori knowledge of the source to reduce the cardinality of the candidate set. Simulation results show that this approach leads to superior performance for moderately skewed sources with linear encoding complexity, which meets the low power consumption requirement of applications such as wireless sensor networks and low-complexity multimedia compression. •By imbedding a linear code in binary arithmetic coding, a Slepian–Wolf code is proposed.•The linear code improves the minimum Hamming distance within the candidate set.•The binary arithmetic coding reduces the cardinality of the candidate set.•Simulation results show that this approach leads to superior performance.
ISSN:0165-1684
1872-7557
DOI:10.1016/j.sigpro.2015.04.013