Integrating AI and IoT for the Implementation of Skin Cancer Detection in Hospital 4.0

Abstract

The growing worldwide incidence of skin cancer, particularly melanoma, highlights the urgent need for inexpensive and efficient diagnostic tools. An AI and IoT system for real-time skin cancer classification optimized for the digital landscape of Hospital 4.0 is proposed in this article. The implementation of a light MobileNetV2 CNN on an edge computing platform based on Raspberry Pi enables the system to perform rapid local image classification via web interface. The system can ensure secure data transmission, interoperability of Electronic Health Record (EHR), and offline capabilities with both clinical and remote uses. Experimental results demonstrate extremely high diagnostic accuracy (96.4 %) and recall (98.5 %), attesting to the system's performance in the early diagnosis of melanoma. The study identifies the potential of embedded AI solutions to enhance diagnostic performance with the ability to maintain cost-effectiveness, scalability, and patient data confidentiality.