Tracking objects in an indoor environment using 3D models

The paper introduces a new approach of tracking real objects in an indoor environment. This approach uses two cameras and 3D models. The changes of the objects in position and pose could be tracked and simulated in a virtual world using predefined information of the indoor environment and estimated parameters of the objects. Our proposed system automatically achieves these parameters by searching the best-matching set of template contours in the database and edge pixels of the objects in two given images using generalized Hough transform