Analysis of gradient flow of a regularized Mumford-Shah functional for image segmentation and image inpainting

Feng, Xiaobing ; Prohl, Andreas 

ESAIM: Mathematical Modelling and Numerical Analysis , Tome 38 (2004) no. 2, pp. 291-320

Résumé

This paper studies the gradient flow of a regularized Mumford-Shah functional proposed by Ambrosio and Tortorelli (1990, 1992) for image segmentation, and adopted by Esedoglu and Shen (2002) for image inpainting. It is shown that the gradient flow with $L^2\times L^{\infty }$ initial data possesses a global weak solution, and it has a unique global in time strong solution, which has at most finite number of point singularities in the space-time, when the initial data are in $H^1\times H^1\cap L^{\infty }$. A family of fully discrete approximation schemes using low order finite elements is proposed for the gradient flow. Convergence of a subsequence (resp. the whole sequence) of the numerical solutions to a weak solution (resp. the strong solution) of the gradient flow is established as the mesh sizes tend to zero, and optimal and suboptimal order error estimates, which depend on $\frac{1}{\epsilon }$ and $\frac{1}{k_{\epsilon }}$ only in low polynomial order, are derived for the proposed fully discrete schemes under the mesh relation $k=o\left(h^{\frac{1}{2}}\right)$. Numerical experiments are also presented to show effectiveness of the proposed numerical methods and to validate the theoretical analysis.

Détail

DOI : 10.1051/m2an:2004014

Classification : 35K55, 65M12, 65M15, 68U10, 94A08
Keywords: image segmentation and inpainting, Mumford-Shah model, elliptic approximation, gradient flow, a priori estimates, finite element method, error analysis