A two-stage procedure using sufficient statistics of the parameters of the source models is proposed. In the procedure, the sufficient statistics calculated from a source sequence is transmitted at the first stage. During the second stage, the source sequence is encoded via the conditional distribution given by the sufficient statistics. The quantization is needed to transmit the estimator vector in the previous two-stage codes. However, because the sufficient statistics is discrete random variable, the quantization is in effect not required to transmit them. Further, the redundancy of the proposed code is equal to that of Bayes code.
|出版ステータス||Published - 1995 1 1|
|イベント||Proceedings of the 1995 IEEE International Symposium on Information Theory - Whistler, BC, Can|
継続期間: 1995 9 17 → 1995 9 22
|Other||Proceedings of the 1995 IEEE International Symposium on Information Theory|
|City||Whistler, BC, Can|
|Period||95/9/17 → 95/9/22|
ASJC Scopus subject areas