An Achievable Rate Region for Distributed Source Coding With Reconstruction of an Arbitrary Function of the Sources
Source: University of Michigan
A new rate region is presented for a very general framework of distributed source coding where the decoder is interested in lossy reconstruction of an arbitrary function of the sources. The coding scheme involves vector quantization of the sources followed by "Correlated" binning. Nested linear codes are used for quantization and binning as against the more prevalent random codes. The authors' rate region recovers many known rate regions in distributed source coding while unifying them under the same general framework.