AI-Based VR Earthquake Simulator

We propose a novel AI-based VR earthquake simulator that can easily simulate an arbitrary real indoor environment. The user first scans any room using inexpensive RGBD sensors, such as Kinect sensors. The automatic model generator then detects objects in RGBD data using a deep-learning-based classification method and generates 3D models given appropriate physical characteristics, such as centroid and friction. In this manner, users can obtain appropriate content for earthquake simulation simply by scanning the room. The content can provide the user with a real experience in the event of an earthquake in the room.