Novel UEP Product Code Scheme with Protograph-Based Linear Permutation and Iterative Decoding for Scalable Image Transmission

This paper introduces a linear permutation module before the inner encoder of the iteratively decoded product coding structure, for the transmission of scalable bitstreams over error-prone channels1.1.A longer version [1] of this work has been submitted to IEEE Transactions on Communications.. This can improve the error correction ability of the inner code when some source bits are known from the preceding outer code decoding stages. The product code consists of a protograph low-density parity-check code (inner code) and Reed-Solomon (RS) codes of various strengths (outer code). Further, an algorithm relying on protograph-based extrinsic information transfer analysis is devised to design good base matrices from which the linear permutations are constructed. In addition, an analytical formula for the expected fidelity of the reconstructed sequence is derived and utilized in the optimization of the RS codes redundancy assignment. The experimental results reveal that the proposed approach consistently outperforms the scheme without the linear permutation module, reaching peak improvements of 1.98 dB and 1.30 dB over binary symmetric channels (BSC) and additive white Gaussian noise (AWGN) channels, respectively. A longer version [1] of this work has been submitted to IEEE Transactions on Communications.