A Skin Type Classification Method Using Mobile Device-Based Deep Learning Model

Skin care products should be tailored to suit different skin types. However, skin testing can be expensive and time-consuming, particularly for students or office workers who may need access to specialized equipment. In the present study, we developed a skin type detection system by using computer-vision and deep-learning techniques that can be easily accessed through a mobile phone application. Our system integrates with the TensorFlow Lite framework on the Android platform and therefore supports various hardware accelerations and easy model validation. TensorFlow Lite, an open-source library developed by Google, is a lightweight, cross-platform, machine-learning framework for mobile and Internet of Things devices. It also supports various hardware accelerations. Our experimental results reveal that the proposed method has an accuracy of 96% and is easy to use on mobile devices. This system provides a convenient and cost-effective means of identifying the skin type and selecting appropriate skin care products.