文章基本信息

标题：Restoration Process for Gray-scale Patterns by Using Path-integral
本地全文：下载
作者：Katsuki KATAYAMA
期刊名称：Interdisciplinary Information Sciences
印刷版ISSN：1340-9050
电子版ISSN：1347-6157
出版年度：2006
卷号：12
期号：1
页码：19-31
DOI：10.4036/iis.2006.19
出版社：The Editorial Committee of the Interdisciplinary Information Sciences
摘要：We investigate a restoration process for a gray-scale pattern given by a snapshot of the Q-states ferromagnetic Husimi–Temperly model within a framework of a Bayesian inference. By using a generating function of path-integral representation, we derive an equation which describes the restoration process for the gray-scale pattern. We investigate two systems: system 1 in which the gray-scale value takes on {−1,−1+²/_{( Q −1)},···,1−²/_{( Q −1)},1} and system 2 in which the gray-scale value takes on {1,···, Q }. We find that there exists difference between Q =2 and Q ≥3 as for behavior of the restoration process for the system 1.
关键词：Bayesian inference;gray-scale pattern;dynamics;generating function of path-integral representation;hyperparameter