3D face recognition
Guide to Biometric Reference Systems and Performance Evaluation, Springer, ss.263-295, 2009
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2009
- Doi Numarası: 10.1007/978-1-84800-292-0_9
- Yayınevi: Springer
- Sayfa Sayıları: ss.263-295
- Boğaziçi Üniversitesi Adresli: Evet
Özet
Three-dimensional human facial surface information is a powerful biometric modality that has potential to improve the identification and/or verification accuracy of face recognition systems under challenging situations. In the presence of illumination, expression and pose variations, traditional 2D image-based face recognition algorithms usually encounter problems. With the availability of three-dimensional (3D) facial shape information, which is inherently insensitive to illumination and pose changes, these complications can be dealt with efficiently. In this chapter, an extensive coverage of state-of-the-art 3D face recognition systems is given, together with discussions on recent evaluation campaigns and currently available 3D face databases. Later on, a fast Iterative Closest Point-based 3D face recognition reference system developed during the BioSecure project is presented. The results of identification and verification experiments carried out on the 3D-RMA database are provided for comparative analysis. © 2009 Springer-Verlag London.