Source-controlled channel decoding in image transmission

Abstract

With the fast growing importance of wireless image communication, the need to deal with fading and other channel distortions arises, in order to enhance the overall throughput and quality of transmitted images. Besides the classical way of combined source and channel coding (Modestino and Daut, 1979) we present a new approach by using source-controlled channel decoding (Hagenauer, 1995) that can be applied either in addition or in conjunction with the above mentioned method. Since there is no optimum source encoder, the generated source data still contains some amount of redundancy (Ruf, 1995). An extension of the well known Viterbi algorithm (Forney, 1973), the a priori soft output Viterbi algorithm (APRI-SOVA), uses the remaining redundancy to improve the error correcting capability. This leads to gains in picture quality of up to 2.2 dB (PSNR) or up to 0.5 dB in channel signal-to-noise ratio (E/sub s//N/sub 0/). Furthermore, the APRI-SOVA allows a more graceful degradation in the case of degrading channel conditions when used for channel decoding instead of conventional Viterbi decoding in image transmission.