Summary
Presented: 2024 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), Pacific Sutera Hotel, Kota Kinabalu, Sabah, Malaysia Cervical cancer is a significant cause of death around the world. In 2022, 660,000 individuals have been diagnosed with the disease around the world and 350,000 individuals have died as a result of it. Current methods of diagnosis require pathologists to look at the cells under a microscope, which is subjective and time-consuming. We present a deep learningbased detection method to automatically classify four different stages (High squamous intra-epithelial lesion, Negative for intraepithelial malignancy, Low squamous intra-epithelial lesion, and Squamous cell carcinoma) of cervical cancer. In this study, we choose to experiment with a lightweight model like the MobileNetV2 model which can be deployed in hardware-constrained scenarios, to evaluate its performance in the preliminary diagnosis of cervical cancer. Our model achieved a validation accuracy of 98.95% and a test accuracy of 94.68%. These results demonstrate the potential effectiveness of the MobileNetV2 model in aiding the early detection of cervical cancer, highlighting its applicability in medical diagnosis tasks.