An Automatic Vehicle Occupant Counting Algorithm Based on Face Detection

In this paper the method for automatically detecting and counting vehicle occupants is studied and a face detection algorithm is proposed to count vehicle occupants. A lighting compensation skin tone model is employed as coarse detection. The correlation of human skin colors is used as fine detection. And mathematical morphological operations are employed. Those three steps detect skin regions of the interiors of vehicles and then generate the face candidates. Finally, a fuzzy ART classifier is employed to perform the occupant detection and counting function. Experimental results show that the algorithm is effective under normal daytime conditions, even in the cases of weak luminance on faces and strong light reflections on the front window