A novel approach to localize human iris centers in facial images is proposed. From the estimated nose center of the detected face, the proposed method searches for the possible iris edge by gradually growing a rectangle, and then localizes the iris center in an expected region by searching for the pixel, which has the minimum average intensity. Experimental results show that this method has good performance for facial images under variations in pose and lighting direction. The speed of the method is notably fast. It has been successfully applied to a face recognition system for providing the eye locations for aligning faces.