Tracking People by Using Color Information from Omnidirectional Images

Abstract

This paper proposes a method for finding and tracking people in input images which uses skin color information. Making color models for skin color and background enables us to find people easily and to track people without any discomfort. Our proposed method can track overlapped people by considering their mutual positions and textures. We use a omnidirectional image sensor for taking images in order to capture many people for a long time in this paper. We conduct experiments for evaluation of the proposed method.