Dynamics of the maximum marginal likelihood hyperparameter estimation in image restoration: Gradient descent versus expectation and maximization algorithm

Inoue, Jun-ichi; Tanaka, Kazuyuki

PHYSICAL REVIEW E, 2002, 65(1), 016125

Permalink : https://hdl.handle.net/2115/5509

このアイテムのアクセス数:2,053件（2026-03-18 00:24 集計）

ファイル	フォーマット	サイズ	閲覧回数	説明
PRE65-1	pdf	689 KB	1,541

アクセス権	open access
URI	https://hdl.handle.net/2115/5509
タイトル	Dynamics of the maximum marginal likelihood hyperparameter estimation in image restoration: Gradient descent versus expectation and maximization algorithm
著者	e-Rad_Researcher 30311658 著者名 Inoue, Jun-ichi 井上, 純一
著者	著者名 Tanaka, Kazuyuki
言語	英語
発行日	2002-01
出版者	The American Physical Society
収録物名	PHYSICAL REVIEW E
巻(号)	65(1)
ページ	016125
抄録	Dynamical properties of image restoration and hyperparameter estimation are investigated by means of statistical mechanics. We introduce an exactly solvable model for image restoration and derive differential equations with respect to macroscopic quantities. From these equations, we evaluate relaxation processes of the system to the equilibrium state. Our statistical mechanical approach also enables us to investigate the hyperparameter estimation by means of maximization of the marginal likelihood by using gradient descent and the expectation and maximization algorithm from the dynamical point of view.
資源タイプ	学術雑誌論文
出版タイプ	Version of Record
権利情報	Copyright © 2001 American Physical Society
PISSN	1063-651X
関連情報	PMID 11800754
関連情報	URI http://www.aps.org/
関連情報 (isIdenticalTo)	出版社版DOI https://doi.org/10.1103/PhysRevE.65.016125