Computing the 3D face recognition based on pseudo 2D Hidden Markov Models using geodesic distances

Face recognition is a very challenging issue and has attracted much attention over past decades. First a 3D mesh model of face is manipulated to calculate the geodesic distances on triangulated surface by the dasiafast marching on triangulated domainspsila (FMTD) numerical algorithm. Then these are processed by multidimensional scaling in order to embed the surface into a low dimensional space. Next, by using the so constructed canonical forms and a 3-6-6-6-3 structure of pseudo 2-D hidden Markov models, an interesting face recognition is carried out.