A Quasi Random Symbol Interleaving Technique Applied to Image Transmission By Noisy Channels

Image vector quantization (VQ) is sensitive to channel errors because they can lead to spurious blocks and consequent quality degradation of transmitted images. One way to increase the robustness of VQ to channel errors is to perform an index assignment to the codevectors. For wireless channels with Rayleigh fading, another challenge is the occurrence of burst errors. This work proposes the application of low discrepancy sequences to the interleaving of symbols, to improve the quality of reconstructed images considering those errors. Results obtained by preference tests, based on visual inspection of the reconstructed images, attest the success of the technique.