A Fully Global Approach to Image Segmentation via Coupled Curve Evolution Equations

By Anthony J. Yezzi Jr., Andy Tsai, Alan S. Willsky. Journal of Visual Communication and Image Representation, 2002

356 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition, Modeling and Simulation

View Paper: https://www.semanticscholar.org/paper/5688cf0d81e0ac7536eeba7dabed3c3fd501a215

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

In Modeling and Simulation

Citations in the corpus

Curve evolution implementation of the Mumford-Shah functional for image segmentation, denoising, interpolation, and magnification (2001, methodology): “In contrast to most other region based techniques however (including our own previous work [48], [49] and that of Chan-Vese [9] and Paragios-Deriche [27]), which assume highly constrained parametric models for pixel intensities within each region, our approach employs the statistical model directly ”
On the relationship between parametric and geometric active contours (2000, background): “Modifications and enhancements have been added to change their behavior or improve their performance in a variety of applications [6, 7, 17–19], including a number of more global region based models which have appeared recently in the literature [35–40].”
Geodesic Shooting for Computational Anatomy (2006, background): “This second point of view in terms of the conservation of momentum law also sheds new light on a large number of high dimensional evolution based Active Model Methods in Computer Vision, including active snakes and contours [6, 12, 18, 20, 29, 31, 37, 40, 41, 43], active surfaces and deformable mode”
Spatially adaptive techniques for level set methods and incompressible flow (2006, background): “For example, [12] (see also [103]) proposed a level set model for active contours to detect objects whose boundaries are not necessarily defined by a gradient.”
Localizing Region-Based Active Contours (2008, unknown): “Some of the most well-known and widely used region-based active contour models assume the various image regions to be of constant intensity [13]–[16].”