Skin Care Product Recommendation Using Convolutional Neural Networks

Abstract

This study proposes a novel approach for skin care product recommendation through Convolutional Neural Networks (CNNs). Leveraging various deep learning architectures including ResNet50, CNN, InceptionV3, DenseNet201, and Xception, our classification model achieved significant improvements in accuracy compared to previous methods. Specifically, the Xception model yielded remarkable results with a classification accuracy of 99% on the Skin Type dataset. Additionally, we explored object detection techniques using YOLO models (versions V5, V6, V7, V8) to identify and localize skin disease regions, thus enhancing the robustness and comprehensiveness of our analysis. This research contributes to the advancement of skin care recommendation systems by harnessing the power of state-of-the-art deep learning architectures for both classification and detection tasks. Our findings underscore the potential of CNNs in revolutionizing personalized skin care solutions, ultimately leading to more effective and targeted skincare recommendations for individuals with diverse skin types and conditions.