Low cost head tracking system for a desktop-based VR system using webcam

With the advancement of computer technology, vision-based system has found various applications ranging from video surveillance, object recognition, industrial defect inspection and autonomous robots. The fundamental part for many computer vision tasks is object detection. This paper discusses the development of a low cost head tracking system for a Desktop-based VR System using a webcam. OpenCv' OpenGL and C/C++ are employed as the development tools for video acquisition, face detection and a virtual environment. At the first stage, a video stream of the webcam is captured. The face is detected by using the Haar-like feature method. Then, a computer vision technique is used to track the position of the human head in sequence of the frames of the video stream. The tracking data is filtered and sent to a VR system in real-time, in which the position of the objects is moved according to the human head in the video stream of a webcam. Time multiplex technique is used to create the stereoscopic images and viewed by shutter glasses. Collected data is compared with video analyzing method and commercial available magnetic sensors. Results obtained demonstrate the effectiveness of the system in face detection with some limitation such as detecting the face in darkness. The tracking system is only able to detect 2D positions, in which depth cannot be process accurately