Point Distribution Models for Pose Robust Face Recognition: Advantages of Correcting Pose Parameters Over Warping Faces to Mean Shape

Daniel González-Jiménez, José Luis Alba-Castro

2007

Abstract

In the context of pose robust face recognition, some approaches in the literature aim to correct the original faces by synthesizing virtual images facing a standard pose (e.g. a frontal view), which are then fed into the recognition system. One way to do this is by warping the incoming face onto the average frontal shape of a training dataset, bearing in mind that discriminative information for classification may have been thrown away during the warping process, especially if the incoming face shape differs enough from the average shape. Recently, it has been proposed a method for generating synthetic frontal images by modification of a subset of parameters from a Point Distribution Model (the so-called pose parameters), and texture mapping. We demonstrate that if only pose parameters are modified, client specific information remains in the warped image and discrimination between subjects is more reliable. Statistical analysis of the verification experiments conducted on the XM2VTS database confirms the benefits of modifying only the pose parameters over warping onto a mean shape.

References

  1. Beymer, D.J. and Poggio, T., “Face Recognition from One Example View,” in Proceedings ICCV 1995, 500-507.
  2. Turk, M. and Pentland, A., “Eigenfaces for recognition,” in Journal of Cognitive Neuroscience Vol. 3, 1991, pp. 7286.
  3. Wiskott, L., Fellous, J.M., Kruger, N., von der Malsburg, C., “Face recognition by Elastic Bunch Graph Matching,” in IEEE Transactions on PAMI, Vol. 19, No.7, 1997 pp. 775-779.
  4. Xiujuan Chai, Shiguang Shan, Xilin Chen, Wen Gao, “Local Linear Regression (LLR) for Pose Invariant Face Recognition,” in Proceedings of 7th International Conference on AFGR'06, 631-636.
  5. Xiujuan Chai, Laiyun Qing, Shiguang Shan, Xilin Chen, Wen Gao, “Pose Invariant Face Recognition under Arbitrary Illumination based on 3D Face Reconstruction,” in AVBPA 2005, NY, USA, 2005, 956-965.
  6. Blanz, V. and Vetter, T., “A Morphable model for the synthesis of 3D faces,” in Proceedings SIGGRAPH 1999, 187-194.
  7. Blanz, V., Grother, P., Phillips, P.J., and Vetter, T., “Face Recognition Based on Frontal Views Generated from Non-Frontal Images,” in Proceedings IEEE Conference on CVPR 2005, 454-461.
  8. Vetter T., and Poggio T., “Linear Object Classes and Image Synthesis from a Single Example Image,” IEEE PAMI vol. 19, pp. 733-742, 1997
  9. González-Jiménez, D. and Alba-Castro, J.L., “Pose Correction and Subject-Specific Features for Face Authentication”, in ICPR 2006, Vol. 4.
  10. Lanitis, A., Taylor, C.J. and Cootes, T.F.,“Automatic Interpretation and Coding of Face Images Using Flexible Models,”in IEEE PAMI, Vol. 19, No. 7, pp. 743-756, 1997.
  11. Messer, K., Matas, J., Kittler, J., Luettin, J., and Maitre, G. XM2VTSDB: “The extended M2VTS database,” in Proceedings AVBPA 1999, 72-77.
  12. Luttin, J. and Maˆitre, G. “Evaluation protocol for the extended M2VTS database (XM2VTSDB),” Technical report RR-21, IDIAP, 1998.
  13. Bookstein, Fred L.: “Principal Warps: Thin-Plate Splines and the Decomposition of Deformations,” in IEEE PAMI, Vol. 11, No. 6, April 1989, pp. 567-585.
  14. Bengio, S., Mariethoz, J. “A statistical significance test for person authentication”, in Proceedings Odyssey 2004, 237-244.
  15. Lyons, M.J., Campbell, R., Plante, A., Coleman, M., Kamachi, M., and Akamatsu, S., “The Noh Mask Effect: Vertical Viewpoint Dependence of Facial Expression Perception,” in Proceedings of the Royal Society of London B 267: 2239-2245 (2000).
Download


Paper Citation


in Harvard Style

González-Jiménez D. and Luis Alba-Castro J. (2007). Point Distribution Models for Pose Robust Face Recognition: Advantages of Correcting Pose Parameters Over Warping Faces to Mean Shape . In Proceedings of the 7th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2007) ISBN 978-972-8865-93-1, pages 138-147. DOI: 10.5220/0002428001380147


in Bibtex Style

@conference{pris07,
author={Daniel González-Jiménez and José Luis Alba-Castro},
title={Point Distribution Models for Pose Robust Face Recognition: Advantages of Correcting Pose Parameters Over Warping Faces to Mean Shape},
booktitle={Proceedings of the 7th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2007)},
year={2007},
pages={138-147},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002428001380147},
isbn={978-972-8865-93-1},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 7th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2007)
TI - Point Distribution Models for Pose Robust Face Recognition: Advantages of Correcting Pose Parameters Over Warping Faces to Mean Shape
SN - 978-972-8865-93-1
AU - González-Jiménez D.
AU - Luis Alba-Castro J.
PY - 2007
SP - 138
EP - 147
DO - 10.5220/0002428001380147