Harnessing Technology for Precise Detection and Classification of Cervical Cancer

Cervical cancer is one of the most prevalent diseases among women and the leading health concern for the increase in the female mortality rate globally. With approximately 500000 new cases reported annually around the world, the early detection and analysis of cancer have become critical. Traditional Screening techniques such as Pap smears, colposcopy, Human Papillomavirus (HPV), and biopsies are well-known screening approaches for the detection of cervical cancer often associated with challenges like time-consuming, costly, invasive, and requiring highly skilled pathologists, which increases susceptibility to human error. Overcoming the challenges of traditional screening approaches, the advancement of medical image processing, and artificial intelligence with innovative solutions have enabled more accurate and efficient cervical cancer detection. This study proposes a hybrid CNN approach for detecting and analyzing cervical cancer cell attributes from papsmear images. An automated tool is designed with a simple graphical user interface to identify cervical cancer and its five types of cancer tissues. The region extraction and feature extraction are implemented using SIPaKMeD and Herlev datasets. Overall, the proposed model demonstrates 89.13% classification accuracy for five cervical cell variants.