Open AccessAnalysis of Image Restoration Based on EM Algorithm
Author(s) -
Yujie Lu
Publication year - 2022
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2242/1/012045
Subject(s) - convergence (economics) , algorithm , gaussian , estimation theory , mathematics , gaussian noise , image (mathematics) , noise (video) , image restoration , mathematical optimization , expectation–maximization algorithm , distribution (mathematics) , maximum likelihood , computer science , image processing , statistics , artificial intelligence , mathematical analysis , physics , quantum mechanics , economics , economic growth
Image noise is mainly introduced in the process of image imaging. Different imaging mechanisms lead to different distribution characteristics of image noise. The maximum likelihood method is a commonly used parameter estimation method, which can obtain a good estimation effect for Gaussian distribution, but it has the disadvantage of relatively complicated calculation when it is used for parameter estimation of mixed Gaussian distribution. However, it has the disadvantage of relatively complicated calculation when it is used for parameter estimation of mixed Gaussian distribution. The EMs of this paper flexibly bring their blurring effects. The solution of the maximum likelihood equation is simplified by introducing potential variables, which has the advantages of smooth implementation of equations, stable numerical calculation, and guaranteed convergence to a stable point to make up for the deficiency of the maximum likelihood method