Extracting human facial baseline from measuring the fractal signature

Automatic orientation and extraction of facial feature is significant in many practical applications. In this paper, we present a new method for extracting the face baseline from measuring the fractal signature. The procedure of the detection is: segment the face out of the background; compute the fractal signature with a sampling window over the facial region; search two peaks in the upper part of the fractal signature graph, the connection of the two peaks are defined as the face baseline we are looking for. The detection method has a strong robustness for object rotation, noise, and different lighting. The result of experiments shows the method is valid for the detection of watermark in the scanning of person identification cards.