Skip to Main Content
In this paper, we propose a novel hybrid method for the segmentation and automatic counting of biological cell image. The method is based on techniques of morphology, thresholding and watershed. It performs well in low contrast image where gradient-based method may fail. Experimental results on practical cell images are shown in the paper with the emphasis on the comparisons between the novel hybrid method and the gradient-based methods: Sobel , Canny  and GAC  of level-set.