Pap-smear image segmentation and stage detection of cervical cancer using deep learning.

Cancer is abbreviated as a generic term for all the diseases. This disease can be fatal if not detected at theproper time and if the diagnosis is inaccurate. The mortality rate due to cancer can be decreased if it is detected at an early stage. This paper takes into consideration the clinical analysis of the Pap-Smear results of 30 patients and uses deep learning algorithms to segment the overlapping cell with the nucleus. As U-Net architecture requires less images, so the proposed algorithm can work on a smaller number of images. Then the Support Vector Machine classifier is used to detect the stages of the cancer. For segmenting Bio medical images U-Net and Region Proposal Network is used. In the proposed algorithm we leverage the U-Net with the RPN. The training of both the network is done separately rather it is done together. The output of the combined network which are segmented cells and nucleus of cells are put into the SVM classifier for the detection of the stages of the cancer. The false positive rate, false negative rate, true positive rate and true negative rate of the proposed algorithm is compared with another existing algorithm. The comparison results are included in the end of the paper in a separate section. [ABSTRACT FROM AUTHOR]