Low complexity technique exploiting 2D source correlation using single parity check codes

This paper proposes a low complexity joint source-channel coding (JSCC) technique that can well exploit two-dimensional (2D) source correlation using turbo single parity check (SPC) codes. Based on similar structure of the product accumulate code, a rate-1 inner code is serially concatenated to the turbo SPC code structure in order to overcome the high error floor issue of turbo SPC codes. The 2D source correlation, horizontal and vertical directions of the source are exploited during the decoding process. In order to exploit the source statistics (i.e. transition probabilities), a decoding algorithm based on modified check operation is proposed. This modified check operation has significantly lower computational complexity than the previous technique that based on modified Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm. Simulation results reveal that the proposed technique that based on modified check operation achieves about 4.5 times lower computational complexity than the technique that based on modified BCJR algorithm. In addition to the lower computational complexity, the proposed JSCC technique can achieve optimal performance, similar with the technique that based on modified BCJR algorithm. The interesting features of low computational complexity and high performance of the proposed technique is important in designing JSCC system exploiting higher dimensional source correlation.