ABSTRACT
This chapter presents a comprehensive framework for artificial intelligence (AI)-driven skin cancer identification, integrating deep learning, explainable AI (XAI), and public health perspectives. It emphasizes AI’s transformative potential in early detection, risk prevention, and equitable healthcare delivery. Using datasets such as HAM10000 and PH2, the proposed system employs a Modified SegNet for precise lesion segmentation and an Attention-Based Convolutional Neural Network (ABCNN) for classification, achieving an accuracy of 98.6%. Explainability is ensured through Grad-CAM, LIME, and Layer-wise Relevance Propagation, promoting clinician trust and ethical transparency. The chapter also details a Python-based implementation and deployment through web and mobile applications for real-time diagnosis. By combining public health intelligence, preventive monitoring, and advanced AI methodologies, this framework demonstrates how intelligent systems can enhance diagnostic precision, accessibility, and inclusivity in dermatological care.
