Classification of Cervical Cancer using ResNet-50

Convolutional neural networks as a part of deep learning is one of the best ways to get accurate results in terms of identification and classification of images in the medical field. In terms of public health, cervical cancer is a major issue and early detection and prevention can save lives. Pap smear tests are widely used for cervical cancer screening, where cells from the cervix are collected and examined for abnormal growth. However, accurately analyzing pap smear images can be challenging, and there is a need for automated methods for efficient and accurate diagnosis. In this study, we advocate a classification method for cervical cancer based on pap smear images using a Convolutional Neural Network (CNN) with ResNet50 architecture and applied on the SIPAKMED pap-smear image dataset. The use of ResNet50 architecture in the CNN allows to improve the feature extraction as well as classification. The accuracy levels reached were 97.5%, making it a powerful tool in medical image analysis. The proposed scheme is compared with VGG 11 architecture accuracy of 92%.