13-09-2017, 01:33 PM
Distributed source coding principles or Slepian-Wolf encoding, used primarily for compression of sources correlated with lateral information, may also be used to reduce the complexity of the encoder from a single source. This effectively changes the complexity of the encoder to the decoder. This exchange of computational complexities has opened the door to many applications in wireless video and image communications. So far, most research efforts are concentrated around the compression of correlated sources that are evenly distributed. But nonuniform sources are not infrequent in the real world. In this work we have studied and demonstrated the suitability and optimality of LDPC syndrome codes for distributed source coding of nonuniform sources using Monte Carlo simulation. Our study reveals that the ability of several who approach LDPC codes, in fact approach the Slepian-Wolf bound for non-uniform sources as well. The results of the Monte Carlo simulation show that highly biased sources can be compressed at 0.049 bits / sample away from Slepian-Wolf for moderate block lengths.