Cell segmentation from phase-contrast images using hybrid watershed and region growing algorithm for genomic drug discovery
Jo Orikawa, Toshiyuki Tanaka · Society of Instrument and Control Engineers of Japan · 2010
Currently the existing computer analysis system for genomic drug discovery does not take it in to account that when cells are examined microscopically, excitation light or phosphor that is used on this occasion affects them adversely. We focused attention on phase microscope which can observe cells noninvasively. Until now, phase-contrast image has been difficult to process because recognition rate varies according to conditions, and for this, appropriate analyzing system has not been established. The study succeeded location of almost all cells. Moreover, we could segment 70 or 80 percent of those cells using hybrid watershed and region growing algorithm.