Oral Cancer Detection using Machine Learning

Abstract

Oral cancer is a dangerous health issue that touches thousands of lives annually, particularly in areas with poor access to healthcare centers. Early detection is critical to boosting survival rates, yet numerous individuals lack access to early diagnosis. In this research, we present a smart, light-weight solution through the application of MobileNetV2, a deep learning algorithm with high efficiency and speed. We have developed a simple web application wherein users can input photos of the oral cavity, and the system will immediately tell if there are cancer signs. With the utilization of a pre-trained MobileNetV2 model, our method provides accuracy in addition to running without any issues on low-power devices such as smartphones. The aim is to develop a useful device that can assist physicians and health workers—particularly in remote or underserved communities—and aid in the early detection of oral cancer, when it is most curable.