Early Skin Cancer Detection Based on MobileNet & VGG-16

Abstract

Skin cancer poses a significant global health risk, necessitating early detection to improve patient outcomes. This paper presents a novel approach for early skin cancer detection by integrating MobileNetV3 Small and VGG-16 neural networks. MobileNetV3 Small is optimized for mobile applications, achieving an 80% accuracy rate, while VGG-16, used in cloud-based analysis, enhances diagnostic precision with an 82% accuracy. This dual-model strategy combines the efficiency of MobileNet for on-device classification and the robustness of VGG-16 for detailed server-based analysis, thus advancing AI-driven skin cancer screening. Our results demonstrate the potential of this approach to significantly improve early detection rates, thereby enhancing treatment success and patient survival.