The Insight Journal logo

Implementation of weighted Dijkstra’s shortest-path algorithm for n-D images

Weizman, Lior, Freiman, Moti, Joskowicz, Leo
Hebrew University of Jerusalem
The Insight Journal logo

Please use this identifier to cite or link to this publication:
New: Prefer using the following doi:
Submitted by Lior Weizman on 2009-01-29 15:07:57.

This paper describes the ITK implementation of a shortest path extraction algorithm based on graph representation of the image and the Dijkstra shortest path algorithm. The method requires the user to provide two inputs: 1. path information in the form of start, end, and neighboring mode, the form of which path is allowed to propagate between neighboring pixels, and 2. a weighting function which sets the distance metric between neighboring pixels. A number of perspectives for choosing weighting functions are given, as well as examples using real images. This paper can also serve as an example for utilizing the Boost C++ graph library into the ITK framework.