Curvature based human face recognition using depth weighted Hausdorff distance

In this paper, we propose a novel implementation of a person verification system based on depth-weighted Hausdorff distance (DWHD) using the surface curvatures of the human face. This new method incorporates the depth information and curvatures of local facial features. The weighting function used in this paper is based on depth values, which have differential properties of a face according to the people, so that the distance of this extracted edge maps are emphasized. Experimental results based on the combination of the maximum, minimum, and Gaussian curvature according to threshold values show that DWHD achieves recognition rate of 92.8%, 97.6% and 92.8% of the cases for 5 ranked candidates, respectively, and the proposed method of combined recognition rate for each curvature shows the best.