Shape and texture based face recognition on overlapping 3D sub-regions

Abstract

In this study, we present a new method for handling facial expression variations by dividing face into overlapping 3D sub-regions and extracting independent features by using 3D DCT from each sub-region. By the proposed method, 99.66% rank-one recognition rate is achieved using FRGC ver2.0 database. © 2011 Springer Science+Business Media B.V.