Construction of universal data compression schemes for sensor network

点击翻译按钮获取中文摘要

A condition for the existence of a universal Slepian-Wolf code for general sources is investigated. Our results show the following fact: Assume that we can construct a universal code for a set of general correlated sources, provided that not only the decoder but the encoder also has access to the side information. Then, even when the side information is not available to the encoder, we can construct a universal Slepian-Wolf code for the set of sources. Moreover, we also investigate the trade-off between the error probability and the coding rate in finite regime. Further, we investigate the Heegard-Berger coding problem under the logarithmic distortion measure. Especially, our results give a tight bound for the achievable rate-distortion region for a special case known as the complementary delivery coding problem.