A Fast Optimization Method for Level Set Segmentation

Authors:
Thord Andersson;Gunnar Läthén;Reiner Lenz;Magnus Borga
Affiliations:
Department of Biomedical Engineering, Linköping University, and Center for Medical Image Science and Visualization (CMIV), Linköping University,;Department of Science and Technology, Linköping University, and Center for Medical Image Science and Visualization (CMIV), Linköping University,;Department of Science and Technology, Linköping University, and Center for Medical Image Science and Visualization (CMIV), Linköping University,;Department of Biomedical Engineering, Linköping University, and Center for Medical Image Science and Visualization (CMIV), Linköping University,
Venue:
SCIA '09 Proceedings of the 16th Scandinavian Conference on Image Analysis
Year:
2009

Citing 5
Cited 0

Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations

Journal of Computational Physics
A PDE-based fast local level set method

Journal of Computational Physics
Signal Processing for Computer Vision

Signal Processing for Computer Vision
Designing Spatially Coherent Minimizing Flows for Variational Problems Based on Active Contours

ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Sobolev Active Contours

International Journal of Computer Vision

Quantified Score

Hi-index	0.00

Visualization

Abstract

Level set methods are a popular way to solve the image segmentation problem in computer image analysis. A contour is implicitly represented by the zero level of a signed distance function, and evolved according to a motion equation in order to minimize a cost function. This function defines the objective of the segmentation problem and also includes regularization constraints. Gradient descent search is the de facto method used to solve this optimization problem. Basic gradient descent methods, however, are sensitive for local optima and often display slow convergence. Traditionally, the cost functions have been modified to avoid these problems. In this work, we instead propose using a modified gradient descent search based on resilient propagation (Rprop), a method commonly used in the machine learning community. Our results show faster convergence and less sensitivity to local optima, compared to traditional gradient descent.