Automatic extraction of contour lines and feature points from profile images

This paper proposes a new extraction algorithm of contour lines and feature points from profile images for the automatic personal identification. First we input a side view of a human head within the dark background under uniform lighting conditions by a CCD camera, then transform the input to a color difference signal image. After enhancing edges of the image by the Sobel operators we binarize the luminance level of each pixel and portray a black profile in a white background. Next we differentiate the black profile, binarize the differentiated results, and depict a contour line using a thinning operator. Finally we encode the extracted contour line by Freeman's chain code and, using the encoded data, calculate the digital curvature in concave sections of the contour line and determine the concave feature points. Next we draw straight lines connecting the adjacent feature points and specify the convex feature points on the contour line which are farthest from the straight lines. We can obtain feature points of the profile automatically and reliably.