Segmentation, recognition and tracing analysis for high-content cell-cycle screening

Abstract

We present in this paper some new and efficient algorithms for segmentation, recognition and tracing analysis of cell phases for high-content screening. The conceptual frameworks are based on the morphological structures of cells where a series of morphological structural points are established. Furthermore, we address the issue of touching cells and then propose morphological techniques for cell separation, reconstruction and tracing analysis. The new segmentation method can resolve the question of over-segmentation. The tracing analysis of cell phases is based on cell shape, geometrical features and difference information of corresponding neighbor frames. Experiment results test the efficiency of the new method