Hypergraph-based Coding Schemes for Two Source Coding Problems under Maximal Distortion

Abstract

We consider two problems in multiterminal source coding under maximal distortion: distributed coding for computing and successive refinement. In distributed coding for computing, we propose a hypergraph-based coding scheme which matches the sum-rate bound of the Berger-Tung inner bound. Further, this scheme matches the entire Berger-Tung inner region when the sources are independent and it outperforms existing graph-based coding schemes. For successive refinement, we propose a hypergraph-based scheme that attains the entire rate region.