Minimum distortion variance concatenated block codes for embedded source transmission

Some of the state-of-art multimedia source encoders produce embedded source bit streams that upon the reliable reception of only a fraction of the total bit stream, the decoder is able reconstruct the source up to a basic quality. Reliable reception of later source bits gradually improve the decoder reconstruction quality. Examples include scalable extensions of H.264/AVC and progressive image coders such as JPEG2000. To provide an efficient protection for embedded source bit streams, a concatenated block coding mechanism using a minimum mean distortion criterion was considered in the past. Although, the original design was shown to achieve mean distortion characteristics better than that of previous research, the proposed coding structure was leading to dramatic quality fluctuations. In this paper, a modification of the original design is first presented and then the second order statistics of the distortion is taken into account in the optimization. More specifically, an extension scheme is proposed using a minimum distortion variance optimization criterion. This robust system design is tested for an image transmission scenario. Numerical results show that the proposed extension achieves significantly lower variance than the original design, while showing similar mean distortion performance using both convolutional codes and low density parity check codes.