Bilkent University
Department of Computer Engineering


Smart Markers for Watershed-Based Segmentation of Living Cells


Can Fahrettin Koyuncu
MSc Student
Computer Engineering Department
Bilkent University

Automated live cell imaging systems facilitate fast and reliable analysis of biological events at the cellular level. In these systems, the first step is usually the cell segmentation, which significantly affects the success of the analysis. On the other hand, similar to other image segmentation problems, the cell segmentation is an ill-posed problem, in which the domain specific knowledge is necessary to obtain successful segmentations even for human subjects. The approaches that can incorporate this knowledge into their segmentation algorithms have a potential to greatly improve the segmentation results.

In this work, a new approach for the effective segmentation of living cells is proposed and developed. This approach introduces a new set of "smart markers" for a marker-controlled watershed algorithm, for which the identification of its markers is critical. The proposed approach relies on using the domain specific knowledge, in the form of visual characteristics of the cells, to define the markers. We evaluated our approach on a total of 1952 cells. The experimental results demonstrated that the incorporation of the domain specific knowledge into the marker definition is quite effective in identifying better markers compared to its counterparts. This would in turn be effective in improving the segmentation performance of a marker-controlled watershed algorithm.

Keywords: cell segmentation, phase-contrast imaging, marker-controlled watershed


DATE: 30 April, 2012, Monday @ 15:40