Automatic segmentation of abnormal cell nuclei from microscopic image analysis for cervical cancer screening

Abstract

In this paper, two methods of microscopic image analysis were presented to classify the abnormal cells in Papanicolaou(Pap) smear for cervical cancer screening. Our goal is to extract those cell nuclei which are abnormally large size, bizarre shape as well as hyper density and we hope to apply this method to the different kinds of abnormal cells. The global information of the image and the local image condition were meantime considered. The reported method searched whole picture by scanning on different axes and determined the locations of abnormal cell nuclei with high contrast This developed method is also able to find cell nuclei, those were almost as bright as the background. Using the reported cytological image analysis, we successfully recognized the abnormal cells such as squamous intraepithelial neoplasia (SIL) and differentiated them from the normal epithelial cells.