3D face recognition based on sparse representation

• DOI: 10.1007/s11227-010-0533-9
• Publication Date: 2010-12-21T14:54:27Z
• Authors: Hengliang Tang; Yanfeng Sun; Baocai Yin; Yun Ge
• Keywords: Normalization; Representation; Three-dimensional face recognition; Feature (linguistics)

In this paper, we present a novel 3D face recognition algorithm based on the sparse representation. First, a 3D face normalization approach is proposed to deal with the raw faces. Then, three types of facial geometrical features are extracted to describe the 3D faces. Meanwhile, in order to guarantee the feasibility of the sparse representation framework and promote the recognition efficiency, a novel feature ranking scheme based on Fisher linear discriminant analysis (FLDA) is designed to arrange the facial descriptors. Finally, the sparse representation framework is used to collect all the face features, and it addresses the recognition task. The experiments tested on the BJUT-3D and FRGC v2.0 databases demonstrate the validity of the proposed 3D face recognition algorithm, and the necessity of the FLDA ranking scheme in the sparse representation framework.