A real-time face detector

In this paper, we present a real-time face detector using a hybrid template that characterizes face-like shapes and face texture patterns. The face detector consists of two main stages: (1) face shape detection based on a modified fast ellipse detection algorithm and (2) face candidate verification in which we define a set of face texture patterns and extend a texture description method to verifying the presence of a face in the face shape candidate region. We present our experimental results which shows that the proposed approach is able to achieve high accuracy with a real-time performance.