A scalable virtual try-on system based on cloud computing

Van Ngoc, D.; Xuan Canh, C.

With the rapid development of artificial intelligence and machine learning technologies, smart devices and smart tools are becoming prevalent. This paper mentions a new technology benefiting the apparel retail industry. In particular, online shopping has developed tremendously for the last decade, especially during Covid-19 pandemic. Numerous fashion and apparel companies are leveraging online channels to boost their sales. However, product fitting (unable to try clothes online) is one of the major challenges that sellers and buyers face in online shopping. Virtual try-on technology has been introduced to tackle this problem. The technology allows shoppers to try garment, jewelry, eyewear, etc on a 3D avatar without wearing these products physically. However, the virtual try-on technology meets some challenges, such as managing a huge amount of data of the 3D model;equipping specialized hardware and securing the privacy of user's biometric data. To address these problems while still ensuring an affordable price as well as the ease of use for customers, a centralized processing system model on cloud computing services is applied. This paper firstly indicates the challenges of deploying cloud systems for virtual try-on solutions, then proposes a well-architecture using a centralized processing system model. © 2021 IEEE.