Adaptive MAP source-controlled channel decoding for MPEG-4 imagery wireless transmission systems

In this paper, we design the adaptive MAP source-controlled channel decoding for transmitting MPEG-4 images. The first order Markov model is employed for adaptively computing the source transition probabilities of the zerotree symbol sequences. Those probabilities are required at the receiver for the newly branch metric computation of the Viterbi algorithm. For analysis, we efficiently evaluate the residual redundancies for both the ldquoLenardquo and the ldquoBarbarardquo images. The system performance is summarized in term of both PSNR (dB) and WER. We consider the slow flat Rician fading channels and also block-fading Rician fading channels. In the simulation results, we achieve the most improvement of about 0.25-0.26 dB for both images. Also, in most cases, the PSNR improvement for the ldquoBarbarardquo image is slightly greater than that for the ldquoLenardquo one.