z-logo
Premium
Robust JPEG image transmission using unequal error protection and code combining
Author(s) -
Fowdur T. P.,
Soyjaudah K. M. S.
Publication year - 2008
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.875
Subject(s) - computer science , network packet , jpeg , transmission (telecommunications) , bit error rate , forward error correction , propagation of uncertainty , channel (broadcasting) , real time computing , algorithm , decoding methods , computer network , data compression , telecommunications
JPEG image transmission over noisy channels is highly problematic due to the sensitivity of the JPEG bit stream to error propagation. The use of resynchronization markers and channel coding do not alleviate the problem completely thus making retransmissions inevitable. In packetized image transmission, image packets are repeated n times, to ensure reliable transmission. This paper proposes a new unequal error protection (UEP) scheme which jointly optimizes the allocation of channel code rates and number of repeats to image packets, subject to a constraint on the maximum overall transmission rate. The coding scheme used is the rate compatible punctured convolutional code coupled with the code‐combining technique. An unequal allocation of headers to the image packets is also performed in order to reduce the overall distortion due to error propagation. Simulation results show that the proposed UEP scheme provides a gain of more than 8 dB in peak‐to‐peak signal‐to‐noise ratio over a tandem scheme. The flexibility of the proposed scheme, and the major performance gains obtained, make the scheme appealing for applications like, web‐based image browsing, multi‐hop networks, and wireless image transmission. Copyright © 2007 John Wiley & Sons, Ltd.

This content is not available in your region!

Continue researching here.

Having issues? You can contact us here