Inertial Nonconvex Alternating Minimizations for the Image Deblurring

Sun, Tao ; Barrio, Roberto (Universidad de Zaragoza) ; Rodriguez, Marcos (Universidad de Zaragoza) ; Jiang, Hao

Resumen: In image processing, total variation (TV) regularization models are commonly used to recover the blurred images. One of the most efficient and popular methods to solve the convex TV problem is the alternating direction method of multipliers (ADMM) algorithm, recently extended using the inertial proximal point method. Although all the classical studies focus on only a convex formulation, recent articles are paying increasing attention to the nonconvex methodology due to its good numerical performance and properties. In this paper, we propose to extend the classical formulation with a novel nonconvex alternating direction method of multipliers with the inertial technique (IADMM). Under certain assumptions on the parameters, we prove the convergence of the algorithm with the help of the Kurdyka-Lojasiewicz property. We also present numerical simulations on the classical TV image reconstruction problems to illustrate the efficiency of the new algorithm and its behavior compared with the well-established ADMM method.

Idioma: Inglés

DOI: 10.1109/TIP.2019.2924339

Año: 2019

Publicado en: IEEE TRANSACTIONS ON IMAGE PROCESSING 28, 12 (2019), 6211-6224
ISSN: 1057-7149

Factor impacto JCR: 9.34 (2019)

Categ. JCR: ENGINEERING, ELECTRICAL & ELECTRONIC rank: 11 / 265 = 0.042 (2019) - Q1 - T1
Categ. JCR: COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE rank: 8 / 136 = 0.059 (2019) - Q1 - T1

Factor impacto SCIMAGO: 2.893 - Software (Q1) - Computer Graphics and Computer-Aided Design (Q1)

Financiación: info:eu-repo/grantAgreement/ES/DGA-FSE/E24-17R
Financiación: info:eu-repo/grantAgreement/ES/MICINN/PGC2018-096026-B-I00
Financiación: info:eu-repo/grantAgreement/ES/MICINN/MTM2015-64095-P

Tipo y forma: Artículo (PostPrint)

Área (Departamento): Área Matemática Aplicada (Dpto. Matemática Aplicada)