Exclusion of photos and new segmentation algorithms for the automatic face recognition

  • Alexander Gebhart
  • Thomas Eppler
Face-based Authentication
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1206)


Face recognition systems have an important disadvantage. It is not always possible to locate the face under the various appearing changes in the surrounding. These changes are above all varying illumination, rotation, translation and size variance. Using special picture processing methods, executed before the verification or identification process is started, these variances can be removed.

The vast majority of face recognition systems work with 2D pictures which were acquired with a standard CCD camera. Analysis has shown, that even the best face recognition systems can be deceived with a photo of a person held in front of the camera. For the exclusion of photos, several alternatives are possible. There are systems which use three dimensional pictures. The heat emission of a face can be measured etc. Nearly all of this methods have on thing in common: they need very complicated and expensive hardware. Easier and more efficient methods that guarantee an exclusion of photos will be presented.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Bichsel, M.: Strategies of Robust Object Recognition for the Automatic Identification of Human Faces. Diss. Nr. 9467, Eidgenössische Technische Hochschule Zürich (ETH) (1991)Google Scholar
  2. 2.
    Danielsson, P.E.: Rotation-invariant Linear Operators with Directional Response. Proc. 5 ICPR (1980)Google Scholar
  3. 3.
    Lades, M., Vorbrügen, J. C., Buhmann, J., Lange, J., von der Malsburg, C., Würtz, R. P., and Konen, W.: Distortion Invariant Object Recognition in the Dynamic Link Architecture. IEEE Transactions on Computers 42(3) (1993) 300–311Google Scholar
  4. 4.
    Seitz, P.: The Robust Recognition of Object Primitives using Local Axis of Symmetry. Elsevier Science Publishers B.V. (1989) 89–108Google Scholar
  5. 5.
    Würtz, R.P.: Multilayer Dynamic Link Networks for Establishing Image Point Correspondences and Visual Object Recognition, Volume 41 of Reihe Physik. Verlag Harri Deutsch, Thun, Frankfurt am Main (1995)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Alexander Gebhart
    • 1
  • Thomas Eppler
    • 1
  1. 1.Eberhard-Karls-Universität TübingenTübingen

Personalised recommendations