Authentication Framework for Augmented Reality with Data-Hiding Technique

Abstract

Currently, most existing research on the security of augmented reality (AR) primarily focuses on user-to-user or user-to-infrastructure authentication, aiming to establish secure communication and prevent potential attacks, such as man-in-the-middle attacks, among others. AR techniques enable users to interact with virtual information and objects within the real environment. However, creating virtual objects can be time-consuming, and unfortunately, malicious individuals may unlawfully copy these objects and integrate them into their own AR scenarios. It is essential to authenticate whether AR objects and the associated content they interact with are legitimately integrated. This paper proposes a novel method for authenticating AR-delivered content using data-hiding techniques with the symmetric property. The proposed method utilizes data hiding to embed data within the content that AR objects interact with, enabling the identification of illegal and tampered AR objects. A scenario involving an AR e-book is defined, and two data-hiding methods are applied to this scenario. The experimental results demonstrate that both data-hiding methods effectively identify the tampered image page of an AR e-book under different tone versions, new trigger image(s) added, or new image(s) replacement.