Distributed multiple description image coding

In this paper a new method for robust image transmission over lossy channels using distributed multiple description coding is presented. We propose an efficient yet a simple method for encoding each description from multiple description scalar quantizer by exploiting the correlation information among descriptions. In this way the correlated information which we call side information can be extracted at the decoder from each description. The motivation to such a type of encoding is to improve the side decoding and to increase the overall robustness of the scheme. Simulation results show that the side distortion is comparable with the central decoding distortion. Performance evaluation of the proposed approach over packet erasure channel shows 0.2-0.7 dB improvement compared to simple multiple description coding that uses non-distributed multiple description scalar quantizer and demonstrates the effectiveness of the scheme over lossy channels.