3D face recognition algorithm based on nose tip contour and radial curve

Human expression variations will cause non-rigid deformation of face scans, this is a challenge in face recognition research. Many rigid registration methods can not achieve good results in dealing with this problem. In this article, a novel method to avoid the influence of non rigid deformation on face recognition which uses radial curve and geodesic contour as features for matching is proposed. It can also reduce data size and speed up calculation. Shape analysis algorithm is also introduced to calculate geodesic distance of corresponding 3D curve of the simplified face as the basis of classification. Experimental results show its high efficiency.