Face Recognition With 3 Dimensional Profile by using Iterative Closest Point Algorithm

In this study, a face recognition method is presented, that is based on best matching after registration of a 3D profile taken from a test face with 3D point cloud of other faces in the database. For 3D registration, iterative closest point algorithm is used. After this operation, the best matching face is chosen according to their distances. Thus, there is no need for early registration of test face and the faces in the database. When there is a new test face, this procedure is applied and the necessary search for matching is done. At the result of experimental studies by using this approach, it is observed that a face can be successfully recognized with only one characteristic profile of it. Thus, it is shown that successful recognition can be done in lack of whole face 3D data. In this approach, the registration process is faster because instead of all points only a profile is used.