Driver’s Face Tracking With Kinect

With the development of social economy, people’s living and working rhythm and intensity are increased. The consequences of the high strength work is working under fatigue state. In this paper, the author’s laboratory has been working on the research of fatigue forecast. And in the research of fatigue based on video, object tracking is an important part.Object tracking has been an important research problem in computer vision. With the publish of Microsoft’s commercial real-time video acquisition device, depth video based object tracking technology become a new research hotspot. This paper presents a sparse coding based object tracking algorithm with depth video, which can solve the problem of illumination changes and occultation in object tracking.Firstly this paper introduces the previous color video based object tracking algorithms, and sparse coding is introduced to solve the problem of partial occultation. In order to reduce the in?uence of the illumination changes, this paper updates the origin object tracking method by introducing depth video, adding Sigmoid normalization and redesigning sparse coding based object tracking algorithm. Finally, in order to solve tracking failure problem during the long-term object tracking, this paper tries the TLD framework and enlarging the search area for object detection. The proposed algorithm can effectively solve illumination changes and occultation problems in object tracking.