An Edge Internet of Things Framework for Machine Learning-Based Skin Cancer Detection Models

Karthik Reddy Kanjula, Ashok Raju Datla, Tianjie Chen, Md Faisal Kabir

Abstract: Skin cancer is one of the most widespread diseases that can be diagnosed through artificial intelligence and computer vision. In recent years, researchers focused on addressing skin cancer at the edge because of enhanced real-time processing capabilities, reduced data vulnerability, and cost-effective hard-ware solutions. Despite the advancements in neural networks and hardware for edge applications, there is still a gap in translating related theoretical findings into practical applications. To bridge this gap, we propose a Internet of Things framework that is lightweight and easily scalable through federated learning. Furthermore, our end-to-end framework could incorporate other CV models and enhance their inference capabilities through edge acceleration. Additionally, we also developed an end-to-end application for mobile devices to detect skin cancer and recommend nearby skin specialists or discussion forums. Our work has paved the road for future machine learning-based edge applications.