Deep Learning-Based Oral Cancer Screening via Smartphone Imagery and Real-Time Web Interface

This study is concerned with the implementation of deep learning techniques, especially Convolutional Neural Networks (CNNs), in picture processing for early oral cancer detection. By combining approaches of oral photos captured through smartphones with image preprocessing methods such as HSV conversion, normalization, and resampling, the algorithm correctly classifies instances as normal or malignant. Early detection has become easier, especially in the underserved or rural areas, due to an intuitive web interface, which allows for real-time image submission and fast diagnosis. A point-by-point comparison of the present-day deep learning and the machine learning solutions used in ascertaining oral cancer is also contained in the paper.