Video-Rate Hair Tracking System Using Kinect

In this paper, we propose automatic hair detection and tracking system that runs at video-rate (30frame per-second) by making use of both the color and the depth information of the images obtained from a Kinect. Our system has three characteristics: 1) Using a 6D feature vector to describe both the 3D color feature and 3D geometric feature of each pixel uniformly; 2) Classifying pixels in images into foreground (e.g. hair) and background with K-means clustering algorithm; 3) Automatic selecting and updating the cluster centers of foreground and background before and during hair tracking. Our system can track hair of any color or style robustly in clustered background where some objects have color similar to the hair, or in environment where the illumination changes. Moreover, our algorithm can be used for tracking a face (or head) if the face (skin+hair) is selected as foreground.