Segmentation and Quantitative Analysis of Normal and Apoptotic Cells from Fluorescence Microscopy Images

Accurate and fast quantitative analysis of living cells from fluorescence microscopy images is useful for evaluations of experimental outcomes and cells culture protocols. An algorithm is developed in this work to automatically segment and discern apoptotic cells from normal cells. A coarse segmentation algorithm is proposed as a pre-filtering step that combines a range filter with a marching square method. This step provides approximate coordinates of cells’ positions in a two-dimensional matrix used to store cells’ image. With this information, the active contours without edges method is applied to identify cells’ boundaries and subsequently it is possible to extract the mean value of intensity within the cellular regions, the variance of pixels’ intensities in the vicinity of cells’ boundaries and the lengths of the boundaries. These morphological features are then employed as inputs to a support vector machine (SVM) classifier that is trained to distinguish apoptotic from normal viable states of cells. The algorithm is shown to be efficient in terms of computational time, quantitative analysis and differentiation accuracy, as compared to the use of the active contours method without the proposed coarse segmentation step.