Region-based segmentation: fuzzy connectedness, graph cut, and other related algorithms


Krzysztof Chris Ciesielski and Jayaram K. Udupa

in Biomedical Image Processing (Thomas M. Deserno, editor), Springer-Verlag, 2011, 251-278.
ISBN 978-3-642-15815-5 e-ISBN 978-3-642-15816-2

In this chapter, we will review the current state of knowledge on region-based digital image segmentation methods. More precisely, we will concentrate on the four families of such algorithms: (i) The leading theme here will be the framework of fuzzy connectedness (FC) methods. (ii) We will also discuss in detail the family of graph cut (GC) methods and their relations to the FC family of algorithms. The GC methodology will be of special importance to our presentation, since we will emphasize the fact that the methods discussed here can be formalized in the language of graphs and graph cuts. The other two families of segmentation algorithms we will discuss consist of (iii) watershed (WS) and (iv) the region growing level set (LS) methods. Examples from medical image segmentation applications with different FC algorithms are also included.

Full text (preprint) in pdf format.

Article full text from Springer (external link)

Entire book full text from Springer (external link)

Last modified April 12, 2011.