Oral Cancer detection using Histopathology Images
Meyya Meyyappan, Aniket Verma, Aarush Kaushik, L Sathyapriya, M. Suresh Anand, Selvanayaki Kolandapalayam Shanmugam · 2024
To increase the chances of survival for the millions of individuals impacted by oral cancer (OC), early detection of the disease is essential. More than 177,384 people died from OC in 2018, and people in the low- to middle-class demographic are primarily affected. The histological examination is the antiquated method of detecting mouth cancer. We are utilizing Deep Learning algorithms to identify OC and automate the process. This paper suggests using histopathology photos to identify OC by building CNN-based models using a transfer learning (TL) technique, then assembling the models to get the best accuracy possible. DenseNet-201, Resnet-50, and EfficientNet-B3 will be used to train the TL models using histopathology pictures. When the three models are combined, the final model has an accuracy of 93.16%, which enables the model to be used in real time and is useful for categorizing images of OC.