Unsupervised segmentation of cervical cell nuclei via adaptive clustering

Abstract

Owing to the uncertainties of manual screening, automated diagnostic tools can aid in improving the reliability of cervical cancer diagnosis. Due to pre-cancerous changes, cell morphology is altered and hence it plays an important role in the screening process. Therefore, segmentation of the cellular parts is important for the classification of cells as normal/abnormal. This paper focuses on segmentation of nuclei in Pap smear images using contrast based adaptive versions of mean-shift and SLIC algorithms followed by an intensity weighted adaptive thresholding. The proposed method is evaluated on Herlev dataset. The performance of the proposed method is compared with state-of-the art clustering based method. The results show that the approach is effective in segmenting images having inconsistent contrast.