A multiphase image segmentation based on fuzzy membership functions and L1-norm fidelity (Q334320)

This paper presents a novel piecewise constant image segmentation model based on fuzzy membership functions and L1-norm fidelity. Observing that the L1-norm is more robust to impulse noise and outliers and can better preserve contrast, in this paper, the authors propose a variational multiphase fuzzy segmentation model based on L1-norm fidelity and fuzzy membership functions. ADMM [\textit{D. Gabay} and \textit{B. Mercier}, Comput. Math. Appl. 2, 17--40 (1976; Zbl 0352.65034)], [\textit{R. Glowinski} and \textit{A. Marroco}, Rev. Franc. Automat. Inform. Rech. Operat., R 9, No. 2, 41--76 (1975; Zbl 0368.65053)] is applied to derive an efficient numerical algorithm, in which each subproblem has a closed-form solution. The proposed method is closely related to the method given in [\textit{Y. He} et al., Pattern Recognition 45, No. 9, 3463--3471 (2012; Zbl 1242.68360)].