The performance of polar codes in distributed source coding

The application of polar coding to rate-adaptive asymmetric Slepian-Wolf coding is considered. The encoder transmits the syndrome set which includes high entropy bits to the decoder to achieve the optimal compression. It means that the encoder must know the correlation between two sources and the syndrome set varies according to this correlation. In this paper, we propose polar source coding scheme with the construction which is designed by density evolution for the target range of crossover probabilities from 0.01 to 0.25. This proposed scheme performs closely to the theoretical Slepian Wolf bound and can be comparable with Low Density Parity Check Accumulate (LDPCA) codes.