A Universal Decoding Scheme for Distributed Lossless Data Compression

摘要

A distributed source coding problem, which includes Slepian-Wolf coding problem as a special case, is considered. A decoding scheme which attains asymptotically the same random-coding error exponent as the maximum-likelihood decoder without knowing source statistics is proposed.%本稿では，Slepian－Wolf符号化問題を特別な場合として含む，情報源の分散符号化問題を考察する．情報源の統計的な性質を利用せずに，最尤復号を用いた場合のランダム符号化誤り指数を漸近的に達成する復号法を提案する．