Automatic Classification of Ovarian Cancer Types Using Cytological Images With the help of DCNN

Abstract

Ovarian cancer is one of the most common gynaecologic tumours and using manual methods it is difficult to detect it, so to ease the hectic process and reduce time required to detect we have developed a process. In this project computer aided diagnosis (CADx) is used which helps the pathologists to determine and to diagnose the tumour correctly. There are basically four types of ovarian cancers.They are serous carcinoma, mucous carcinoma, endometrial carcinoma, clear cell carcinoma. Here cancer types will be detected by using Alex net architecture based on DCNN. In this project, input images areof cytological type. Theimage manipulation is done using different filters to remove noise and obtainsharpened image. There are two types of inputs to DCNN which are training set and testing set. In DCNN there are five convolution layers, three max pooling layers and two full reconnect layers. Input image is passed through DCNN module and respectively type of cancer will be detected.