Momentum Based Optimization Methods for Level Set Segmentation

Authors:
Gunnar Läthén;Thord Andersson;Reiner Lenz;Magnus Borga
Affiliations:
Department of Science and Technology, Linköping University, and Center for Medical Image Science and Visualization, Linköping University,;Department of Biomedical Engineering, Linköping University, and Center for Medical Image Science and Visualization, Linköping University,;Department of Science and Technology, Linköping University, and Center for Medical Image Science and Visualization, Linköping University,;Department of Biomedical Engineering, Linköping University, and Center for Medical Image Science and Visualization, Linköping University,
Venue:
SSVM '09 Proceedings of the Second International Conference on Scale Space and Variational Methods in Computer Vision
Year:
2009

Citing 10
Cited 1

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

Journal of Computational Physics
Learning internal representations by error propagation

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
A PDE-based fast local level set method

Journal of Computational Physics
Signal Processing for Computer Vision

Signal Processing for Computer Vision
Computing Geodesics and Minimal Surfaces via Graph Cuts

ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
Gradient Vector Flow Fast Geometric Active Contours

IEEE Transactions on Pattern Analysis and Machine Intelligence
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
Fast geodesic active contours

IEEE Transactions on Image Processing
Multigrid Geometric Active Contour Models

IEEE Transactions on Image Processing

A shape prior constraint for implicit active contours

Pattern Recognition Letters

Quantified Score

Hi-index	0.00

Visualization

Abstract

Segmentation of images is often posed as a variational problem. As such, it is solved by formulating an energy functional depending on a contour and other image derived terms. The solution of the segmentation problem is the contour which extremizes this functional. The standard way of solving this optimization problem is by gradient descent search in the solution space, which typically suffers from many unwanted local optima and poor convergence. Classically, these problems have been circumvented by modifying the energy functional. In contrast, the focus of this paper is on alternative methods for optimization. Inspired by ideas from the machine learning community, we propose segmentation based on gradient descent with momentum. Our results show that typical models hampered by local optima solutions can be further improved by this approach. We illustrate the performance improvements using the level set framework.