A Person Identification Method Using a Top-View Head Image from an Overhead Camera

In this paper, we describe a novel imagebased person identification task. Traditional facebased person identification methods have a low tolerance for occluded situation, such as overlapping of people in an image. We focus on an image from an overhead camera. Using the overhead camera reduces a restriction of the installation location of a camera and solves the problem of occluded images. First, our method identifies the person’s area in a captured image by using background subtraction. Then, it extracts four features from the area; (1) body size, (2) hair color, (3) hairstyle and (4) hair whorl. We apply the four features into the AdaBoost algorithm. Experimental result shows the effectiveness of our method.