Comparison of segmentation of CU-Net,DRU-Net and DU-Net on Cervical Cancer Data Set

Cervical cancer is one of the most prevalent disease in women.Timely detection of cervical cancer can save lives.Automated detection of cervical cancer tools can be devised whose reliability depends on the rate of accuracy of segmentation of Pap Smear images.For segmenting medical images there are variants of U-Nets available. There are various papers in this area of research that deals particularly with the segmentation of overlapped nucleus.This paper guides the researchers of medical image processing to select the U-Net that gives a trustworthy result.Condensed U-Net,DU-Net and DRU-Net are recent developments in the field of U-Nets.Data sets that are used in this paper are secondary data set taken from 2018 Data Science Bowl on which these three U-Nets are implemented.The accuracy,precision,and Recall values are tabulated to provide a comparison on the performance of these three U-Nets on a cervical cancer data set.