LocIn: Inferring Semantic Location from Spatial Maps in Mixed Reality
LocIn: Inferring Semantic Location from Spatial Maps in Mixed Reality
复制标题
DOI：
--
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Habiba Farrukh;Reham Mohamed;Aniket Nare;Antonio Bianchi;Z. Berkay Celik
中科院分区：
文献类型：
--
作者：
Habiba Farrukh;Reham Mohamed;Aniket Nare;Antonio Bianchi;Z. Berkay Celik
混合现实（MR）设备捕获用户周围环境的3D空间地图，将虚拟内容集成到他们的物理环境中。在流行的MR平台中实现的现有权限模型允许所有MR应用程序在没有明确许可的情况下访问这些3D空间地图。磁共振应用程序对这些3D空间地图的不受监控的访问对用户的隐私构成了严重的威胁，因为这些地图捕捉了用户环境的详细几何和语义特征。在本文中，我们提出了一种新的位置推断攻击，它利用嵌入在3D空间地图中的这些详细特征来推断用户的室内位置类型。L OC I N开发了一种多任务方法来训练端到端编码器-解码器网络，该网络提取空间特征表示以捕获用户环境的上下文模式。L OC I N利用这种表示来检测3D物体和表面，并将它们整合到一个分类网络中，该网络具有新颖的统一优化功能，可以预测用户的室内位置。我们演示了从三种流行的MR设备收集的空间地图上的lc I N攻击。我们表明，lc I N推断用户的位置类型的平均值为84。1%的准确率。
Mixed reality (MR) devices capture 3D spatial maps of users’ surroundings to integrate virtual content into their physical environment. Existing permission models implemented in popular MR platforms allow all MR apps to access these 3D spatial maps without explicit permission. Unmonitored access of MR apps to these 3D spatial maps poses serious privacy threats to users as these maps capture detailed geometric and semantic characteristics of users’ environments. In this paper, we present L OC I N , a new location inference attack that exploits these detailed characteristics embedded in 3D spatial maps to infer a user’s indoor location type. L OC I N develops a multi-task approach to train an end-to-end encoder-decoder network that extracts a spatial feature representation for capturing contextual patterns of the user’s environment. L OC I N leverages this representation to detect 3D objects and surfaces and integrates them into a classification network with a novel unified optimization function to predict the user’s indoor location. We demonstrate L OC I N attack on spatial maps collected from three popular MR devices. We show that L OC I N infers a user’s location type with an average 84 . 1% accuracy.