3D face recognition method based on medium frequency information of geometry image

Facial expression is a main problem in the field of 3D face recognition. To overcome the deficiency,this paper proposed a novel face recognition method based on Detailed Geometry Image (DGI). The DGI, which stays invariant under face expressions, is obtained from multi-scale wavelet transform of three different face regions: the circular area around the nose, the upper face region and the entire face region. Then, an authentication score is gained by combining values of their DGIs. Extensive experiments carried out on the available database FRGC(Face Recognition Grand Challenge) v2.0 achieve a 95.7% verification rate with 0.1% False Acceptance Rate (FAR). Experimental results approve that our method is promising in dealing with expression distortion in 3D face recognition.