AI-driven tools for non-invasive skin analysis: A study in detecting lentigines and nevi in human skin

Abstract

This study demonstrates that deep learning models, particularly YOLOv4 and Faster R-CNN, can effectively detect and segment facial hyperpigmentation with high accuracy. The integration of classical image processing and a user-friendly GUI makes the system accessible to clinicians and researchers. These results highlight the potential of AI to enhance dermatological diagnostics and support longitudinal skin health monitoring.