Turbo Codes Based Image Transmission for Channels with Both Random Errors and Packet Loss

In this paper, we propose an unequal error protection (UEP) framework for progressive image transmission over wireless networks with both random errors and packet loss. By sophisticatedly interleaving each coded frame, the packet loss is converted into randomly punctured bits in a turbo code. Therefore, error control of channels with both types of distortions is equivalent to dealing with random bit errors only through turbo codes with reduced rates. To optimally allocate the overall transmission rate between the source and channel codes, a genetic algorithm (GA) based method is proposed which not only largely reduces the optimization complexity but also gives results close to the brute force search in tested scenarios. Then, we extend the UEP to a product code scheme by adding Reed-Solomon (RS) code across the frames. An efficient rate allocation method for this product code with the use of interleaving is further presented. The effectiveness of the proposed schemes is demonstrated with extensive simulation results.