Abstract
The cervix is the narrow end of the uterus that connects to the vagina in thefemale reproductive system. Abnormal cell growth in the squamous epitheliallining of the cervix leads to cervical cancer in females. A Pap smear is adiagnostic procedure used to detect cervical cancer by gently collecting cellsfrom the surface of the cervix with a small brush and analyzing their changesunder a microscope. For population-based cervical cancer screening, visualinspection with acetic acid is a cost-effective method with high sensitivity.However, Pap smears are also suitable for mass screening due to their higherspecificity. The current Pap smear analysis method is manual, time-consuming,labor-intensive, and prone to human error. Therefore, an artificialintelligence (AI)-based approach for automatic cell classification is needed.In this study, we aimed to classify cells in Pap smear images into fivecategories: superficial-intermediate, parabasal, koilocytes, dyskeratotic, andmetaplastic. Various machine learning (ML) algorithms, including GradientBoosting, Random Forest, Support Vector Machine, and k-Nearest Neighbor, aswell as deep learning (DL) approaches like ResNet-50, were employed for thisclassification task. The ML models demonstrated high classification accuracy;however, ResNet-50 outperformed the others, achieving a classification accuracyof 93.06%. This study highlights the efficiency of DL models for cell-levelclassification and their potential to aid in the early diagnosis of cervicalcancer from Pap smear images.