Remote Access Mathematics of Computation
Green Open Access

Mathematics of Computation

ISSN 1088-6842(online) ISSN 0025-5718(print)

Request Permissions   Purchase Content 
 

 

A posteriori error control for the binary Mumford-Shah model


Authors: Benjamin Berkels, Alexander Effland and Martin Rumpf
Journal: Math. Comp. 86 (2017), 1769-1791
MSC (2010): Primary 49M25, 53C22, 65D18, 65L20
DOI: https://doi.org/10.1090/mcom/3138
Published electronically: October 20, 2016
Full-text PDF

Abstract | References | Similar Articles | Additional Information

Abstract: The binary Mumford-Shah model is a widespread tool for image segmentation and can be considered as a basic model in shape optimization with a broad range of applications in computer vision, ranging from basic segmentation and labeling to object reconstruction. This paper presents robust a posteriori error estimates for a natural error quantity, namely the area of the non-properly segmented region. To this end, a suitable uniformly convex and non-constrained relaxation of the originally non-convex functional is investigated and Repin's functional approach for a posteriori error estimation is used to control the numerical error for the relaxed problem in the $ L^2$-norm. In combination with a suitable cut out argument, fully practical estimates for the area mismatch are derived. This estimate is incorporated in an adaptive mesh refinement strategy. Two different adaptive primal-dual finite element schemes, a dual gradient descent scheme, and the most frequently used finite difference discretization are investigated and compared. Numerical experiments show qualitative and quantitative properties of the estimates and demonstrate their usefulness in practical applications.


References [Enhancements On Off] (What's this?)


Similar Articles

Retrieve articles in Mathematics of Computation with MSC (2010): 49M25, 53C22, 65D18, 65L20

Retrieve articles in all journals with MSC (2010): 49M25, 53C22, 65D18, 65L20


Additional Information

Benjamin Berkels
Affiliation: AICES Graduate School, RWTH Aachen University, 52062 Aachen, Germany
Email: berkels@aices.rwth-aachen.de

Alexander Effland
Affiliation: Institute for Numerical Simulation, University of Bonn, 53115 Bonn, Germany
Email: alexander.effland@ins.uni-bonn.de

Martin Rumpf
Affiliation: Institute for Numerical Simulation, University of Bonn, 53115 Bonn, Germany
Email: martin.rumpf@ins.uni-bonn.de

DOI: https://doi.org/10.1090/mcom/3138
Received by editor(s): May 19, 2015
Received by editor(s) in revised form: December 8, 2015, and December 17, 2015
Published electronically: October 20, 2016
Article copyright: © Copyright 2016 American Mathematical Society