Face segmentation and detection using Voronoi diagram and 2D histogram

Human face segmentation have drawn much attention over the years and many techniques have been used. Seeing that it has many potential applications in computer science vision and automatic access control system, its research domain has promptly broadened not only in engineer but also in the neuroscience. Particularly, face segmentation is a needed step for face recognition system. In this paper, we are going to present a new approach based on color segmentation by exploiting the Delaunay triangulation generated by Voronoi diagram and some feature points from 2D histogram. As for face detection we have used a distance transformation and selected the region of interest. For this work, we have used Georgia Tech face database. The results are quiet satisfying as the success rate reaches 92% for face detection.