Computer Science · Computer Vision and Pattern Recognition
Human tracking using multiple views
Matei Chiperi, Irina Mocanu, Mihai Trăşcău
See on the mapAbstract
Human tracking plays an important role in providing activity assistance and care for users in smart homes. This paper presents a method for detecting and tracking of a user in a smart home using multiple sensors. We consider one PTZ camera and multiple Kinects in order to offer continuity over the users movement. Thus, the user can be keep inside the frame for most of the time. In this way, we minimize the drawbacks of using only one type of sensor. For example, field of view provided by Kinect sensor is not wide enough to cover the entire room. Also, the PTZ camera is not able to detect and track a person in case of different situations, such as the person is sitting or it is under the camera. The system is proposed for the Ambient Intelligence Laboratory (AmI-Lab) at the University Politehnica of Bucharest, and the design is compatible with the software architecture developed for this laboratory, AmI-Platform. The training and evaluation is operated on custom dataset, extracted from the AmI-Lab test environment.