Using mobile devices to capture photos are very common behaviors in our daily life. With such many photos captured from the members belonging to a social network, [Yin et al. 2014] proposed to utilize the social context from the mobile devices, e.g., geo-tag from the GPS sensor, to help a user to capture better photos via a mobile device. Using the geo-tags of a photo and the analysis of the image content to construct a 3D model of a scene has been developed since the Photo Tourism [Snavely et al. 2006] project. The scene reconstruction scheme proposed by [Snavely et al. 2008] can visualize the photos in a 3D environment according to the photos collected from the social members. In addition, [Szeliski et al. 2013] indicated that it is a natural way to navigate the images from the social media sites in a 3D geo-located context. Therefore, for multimedia visualization with a natural and immersive 3D user experience in a tech-art gallery, we propose a 3D social media browsing system to allow users to use motion-sensing devices to interactively navigate the social photos in a virtual 3D scene constructed from a real physical space.