3D Face Recognition using surface normals

In this proposed work, a fully automatic 3D Face Recognition system across pose is presented, which works successfully on three modern databases namely the Frav3D, GavabDB and the Bosphorus databases. Poses handled in the system are yaw, pitch and roll varying from 0° to ±30° as well as expressions. The feature extraction is by depth face images with variation in depth values of the surface normals and also by KPCA. The system gives high recognition rate of 96.92% in case of GavabDB database, 96.25% in case of Bosphorus face database and 92.25% in case of Frav3D database by surface normals extraction.