Nail Type Recommendation using Deep Learning Models

Hands play a crucial role in human interaction and functionality, essential in activities ranging from basic tasks to complex operations. They are essential to fields like biometrics, ergonomics, healthcare, robotics, and the cosmetics industry, particularly in nail care and aesthetics. Understanding hand types can significantly enhance product development and personalization. Traditional methods rely on manual measurements or complex feature engineering, which are labor-intensive and prone to errors.For this reason, this article proposes a novel approach for classifying hands based on their dimensions using deep learning methods to recommend nail types. In this study, deep learning techniques have been leveraged to automatically classify hands into distinct categories based on palm length, palm width, and middle finger length, and suggest nail types for each hand shape accordingly. A dataset of 2050 images was constructed and annotated for classification. Various Convolutional Neural Network (CNN) architectures, including VGG16, LeNet-5, AlexNet, GoogLeNet, ResNet, DenseNet, and MobileNet, were tested and compared for accuracy. VGG16 emerged as the most successful model, achieving high accuracy in classifying hands into predefined categories. Based on these classifications, the proposed model recommends two nail types for each hand type, from a total of seven different nail types. The outcome of the applied model was assessed using standard metrics, advancing hand classification techniques to offer a robust, automated solution for personalized nail recommendations.