On unequal error protection with low density parity check codes in scalable video coding

In this work, we propose an unequal error protection (UEP) scheme for scalable video coding using low density parity check (LDPC) codes. LDPC codes are renowned for their near-capacity performance. In the proposed framework, a video sequence is coded based on scalable video coding (SVC) H.264/AVC standard document. We divide the SVC coded video stream into different sub-streams in terms of time and quality scalability. These sub-streams are channel-coded using low density parity check (LDPC) codes for robustness against channel impairments. LDPC codes with different coding rates are applied for different sub-streams according to the significance level of each sub-stream in order to provide UEP. We develop a probability model for decoding failure using LDPC codes in order to evaluate the bit error rate under different signal-to-noise ratios and delay constraints for an AWGN channel. Further motivated by optimally allocating the bits to maximize the quality of reconstructed video at the decoder against channel induced corruptions, we propose an optimal bit allocation algorithm (OBAA) that optimally allocates the channel coding rates using a dynamic programming approach. We use the proposed probability model in OBAA in order to obtain rate-distortion characteristics of LDPC coded SVC video. The results achieved clearly establish that the proposed UEP scheme can optimally allocate LDPC codes with different coding rates to different sub-streams that significantly decreases the end-to-end distortion of the system as compared to equal error protection (EEP) scheme resulting in greater system throughputs.