Violet Snell segmentation method for fluorescence microscopy

Datum konání: 06.12.2013
Přednášející: Jan Schier
Odpovědná osoba: Kotera

An integral part of analysis of the fluorescence microscopy images is segmentation of cell nuclei and detection of fluorescence signals. Very often, Otsu segmentation is used, as very simple and an easy-to-understand method. However, this metod easily fails in the context of microscopy images, which are characterised by "sparsiness": the area of the foreground objects is much smaller than the are of background. The methods based on active contours provide much better results with the significant drawback of long convergence.

Recently, an interesting method has been proposed by Violet Snell, a PhD student at the University of Surrey. This method applies some ideas from active contours models to threshold-based segmentation. In the talk, we shall briefly review the principles of the fluorescence microscopy and FISH imaging and introduce this method.