Curvature regularity for region-based image segmentation and inpainting: A linear programming relaxation
- 1 September 2009
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15505499,p. 17-23
- https://doi.org/10.1109/iccv.2009.5459209
Abstract
We consider a class of region-based energies for image segmentation and inpainting which combine region integrals with curvature regularity of the region boundary. To minimize such energies, we formulate an integer linear program which jointly estimates regions and their boundaries. Curvature regularity is imposed by respective costs on pairs of adjacent boundary segments. By solving the associated linear programming relaxation and thresholding the solution one obtains an approximate solution to the original integer problem. To our knowledge this is the first approach to impose curvature regularity in region-based formulations in a manner that is independent of initialization and allows to compute a bound on the optimal energy. In a variety of experiments on segmentation and inpainting, we demonstrate the advantages of higher-order regularity. Moreover, we demonstrate that for most experiments the optimality gap is smaller than 2% of the global optimum. For many instances we are even able to compute the global optimum.Keywords
This publication has 16 references indexed in Scilit:
- Computing hypersurfaces which minimize surface energy plus bulk energyPublished by Walter de Gruyter GmbH ,2012
- Fast Image Inpainting Based on Coherence TransportJournal of Mathematical Imaging and Vision, 2007
- Fast Anisotropic Smoothing of Multi-Valued Images using Curvature-Preserving PDE'sInternational Journal of Computer Vision, 2006
- Algorithms for Finding Global Minimizers of Image Segmentation and Denoising ModelsSIAM Journal on Applied Mathematics, 2006
- Level lines based disocclusionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Epi-convergence of discrete elasticaApplicable Analysis, 2001
- Active contours without edgesIEEE Transactions on Image Processing, 2001
- Image inpaintingPublished by Association for Computing Machinery (ACM) ,2000
- Using dynamic programming for solving variational problems in visionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1990
- Trace inference, curvature consistency, and curve detectionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1989