Deep Learning Approaches for Analysing Papsmear Images to Detect Cervical Cancer

The main reason for leading death in female is cervical cancer. The human Papillomavirus is responsible for all cervical cancer cases. There are several diagnostic procedures to find cervical cancer that includes Pap test, liquid-based cytology, colposcopy and HPV test. Early diagnosis is crucial for successful treatment and improved survival rates. In medical image analysis, methods of deep learning are showing promising results for the identification and grading of cervical carcinoma. The work uses dataset of cervical smear images and three cutting-edge deep learning models like ResNet50V2, InceptionV3, and Xception are applied and analysed for prediction of cervical cancer. The Models are verified using cross-validation and the performance are assessed using metrics like accuracy, precision, recall, and F1 score. According to the analysis, ResNet50V2 shows the highest accuracy. The obtained results imply that without the need for invasive procedures, deep learning techniques have the ability to accurately classify cervical cancer and greatly enhance early diagnosis.