Explainable Screening and Classification of Cervical Cancer Cells with Enhanced ResNet-50 and LIME

High levels of recovery rates are proven to be possible at early detection of cancerous cells through various tests, meticulous monitoring, and previous findings. Pap-smear tests are widely used to obtain cervical cells as they are of low cost and due to their painless diagnosis. The use of Deep Learning techniques to create an automated screening and classification system for cervical cancer cells is becoming more and more popular. Our proposed method uses the SIPAKMED pap-smear imagine dataset and Deep Learning classification algorithms to screen for cervical cancer using ResNet-50. The experiment's findings showed that the suggested classification strategy for the ResNet50 architectural model had the highest accuracy (91.04%), outperforming both the Inception-V3 and MobileNet models. In addition, our experiment is advancing further by integrating Explainable Artificial Intelligence (XAI) into the analysis of test images generated by various models. The incorporation of LIME Explainability aims to enhance the interpretability and transparency of the diagnostic process.