Channel decoding of block codes with a priori information

We propose a complete transmission system for still images. This system comprises a VQ source encoder with a compression rate of 9 and the data is encoded using a BCH block code and mapped on a 16-QAM constellation. Two BCH codes are considered: BCH(29,24,3) and BCH(54,48,3). The parameters of the code book and channel encoder have been optimized for 16-QAM. We use a new joint source-channel decoding algorithm for a block code. As compared to classical channel decoding algorithms (hard decoding and soft decoding), we obtain an improvement of 1 dB over a Gaussian channel and 3 dB over a Rayleigh channel, in terms of coding gain for a BER of 10/sup -5/. The PSNR of the received image exhibits an improvement of 3 dB as compared to a conventional system without joint source-channel decoding at low SNRs. The decoded image presents a good subjective quality even if the residual BER is high.