Semi-Linearized Proximal Alternating Minimization for a Discrete Mumford-Shah Model

被引:19
作者
Foare, Marion [1 ]
Pustelnik, Nelly [1 ]
Condat, Laurent [2 ]
机构
[1] Univ Lyon 1, Univ Lyon, ENS Lyon, CNRS,Lab Phys, F-69342 Lyon, France
[2] Univ Grenoble Alpes, CNRS, Grenoble INP, GIPSA Lab, F-38000 Grenoble, France
关键词
Convergence; Image restoration; Image edge detection; Minimization; Noise reduction; Degradation; Couplings; Segmentation; restoration; inverse problems; nonsmooth optimization; nonconvex optimization; proximal algorithms; PALM; Mumford-Shah; ALGORITHMS; REGULARIZATION; APPROXIMATION; SEGMENTATION; RESTORATION; FRAMEWORK; NONCONVEX;
D O I
10.1109/TIP.2019.2944561
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The Mumford-Shah model is a standard model in image segmentation, and due to its difficulty, many approximations have been proposed. The major interest of this functional is to enable joint image restoration and contour detection. In this work, we propose a general formulation of the discrete counterpart of the Mumford-Shah functional, adapted to nonsmooth penalizations, fitting the assumptions required by the Proximal Alternating Linearized Minimization (PALM), with convergence guarantees. A second contribution aims to relax some assumptions on the involved functionals and derive a novel Semi-Linearized Proximal Alternated Minimization (SL-PAM) algorithm, with proved convergence. We compare the performances of the algorithm with several nonsmooth penalizations, for Gaussian and Poisson denoising, image restoration and RGB-color denoising. We compare the results with state-of-the-art convex relaxations of the Mumford-Shah functional, and a discrete version of the Ambrosio-Tortorelli functional. We show that the SL-PAM algorithm is faster than the original PALM algorithm, and leads to competitive denoising, restoration and segmentation results.
引用
收藏
页码:2176 / 2189
页数:14
相关论文
共 48 条
[1]  
Alicandro R., 1999, Interfaces Free Bound, V1, P17, DOI DOI 10.4171/IFB/2
[2]   APPROXIMATION OF FUNCTIONALS DEPENDING ON JUMPS BY ELLIPTIC FUNCTIONALS VIA GAMMA-CONVERGENCE [J].
AMBROSIO, L ;
TORTORELLI, VM .
COMMUNICATIONS ON PURE AND APPLIED MATHEMATICS, 1990, 43 (08) :999-1036
[3]  
AMBROSIO L, 1992, B UNIONE MAT ITAL, V6B, P105
[4]  
[Anonymous], 1987, READINGS COMPUTER VI
[5]   Contour Detection and Hierarchical Image Segmentation [J].
Arbelaez, Pablo ;
Maire, Michael ;
Fowlkes, Charless ;
Malik, Jitendra .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2011, 33 (05) :898-916
[6]   Proximal Alternating Minimization and Projection Methods for Nonconvex Problems: An Approach Based on the Kurdyka-Lojasiewicz Inequality [J].
Attouch, Hedy ;
Bolte, Jerome ;
Redont, Patrick ;
Soubeyran, Antoine .
MATHEMATICS OF OPERATIONS RESEARCH, 2010, 35 (02) :438-457
[7]   Optimization with Sparsity-Inducing Penalties [J].
Bach, Francis ;
Jenatton, Rodolphe ;
Mairal, Julien ;
Obozinski, Guillaume .
FOUNDATIONS AND TRENDS IN MACHINE LEARNING, 2012, 4 (01) :1-106
[8]   Semi-blind image restoration via Mumford-Shah regularization [J].
Bar, L ;
Sochen, N ;
Kiryati, N .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2006, 15 (02) :483-493
[9]  
Bar L., 2011, HDB MATH METHODS IMA
[10]  
Bar L, 2007, IEEE I CONF COMP VIS, P1410