Advertisement

Computational complexity reduction in eigenspace approaches

  • Aleš Leonardis
  • Horst Bischof
Pattern Analysis
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1296)

Abstract

Matching of appearance-based object representations using eigenimages is computationally very demanding. Most commonly, to recognize an object in an image, parts of the input image are projected onto the eigenspace and the recovered coefficients indicate the presence or absence of a particular object. In general, the process is sequentially applied to the entire image. In this paper we discuss how to alleviate the problems related to complexity. First, we propose to use a focus-of-attention (FOA) detector which is intended to select candidate areas of interest with minimal computational effort. Only at these areas we then recover the coefficients of eigenimages. Secondly, we propose to employ a multiresolution approach. However, this requires that we depart from the standard way of calculating the coefficients of the eigenimages which relies on the orthogonality property of eigenimages. Instead we calculate them by solving a system of linear equations. We show the results of our approach on real data.

Keywords

Matched Filter Multiple Resolution Multiresolution Approach Human Face Recognition Illumination Planning 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    T. W. Anderson. An Introduction to Multivariate Statistical Analysis. Wiley, 1958.Google Scholar
  2. 2.
    D. Beymer and T. Poggio. Face recognition from one example view. In Proceedings of 5th ICCV'95. IEEE Computer Society Press, 1995.Google Scholar
  3. 3.
    H. Bischof. Pyramidal Neural Networks. Lawrence Erlbaum Associates, 1995.Google Scholar
  4. 4.
    U. M. Fayyad, P. J. Smyth, M. C. Burl, and P. Perona. Learning to catalog science images. In S. K. Nayar and T.Poggio, editors, Early Visual Learning, pages 237–268. Oxford University Press, 1996.Google Scholar
  5. 5.
    A. Leonardis and H. Bischof. Dealing with occlusions in the eigenspace approach. In Proc. of CVPR96, pages 453–458. IEEE Computer Society Press, 1996.Google Scholar
  6. 6.
    H. Murase and S. K. Nayar.Image spotting of 3D objects using parametric eigenspace representation. In G. Borgefors, editor, The 9th SCIA, volume 1, pages 323–332, Uppsala, Sweden, June 1995.Google Scholar
  7. 7.
    H. Murase and S. K. Nayar. Visual learning and recognition of 3-D objects from appearance. International Journal of Computer Vision, 14:5–24, 1995.CrossRefGoogle Scholar
  8. 8.
    H. Murase and S.K. Nayar. Illumination planning for object recognition using parametric eigenspaces. IEEE Trans. on PAMI, 16(12):1219–1227, 1994.Google Scholar
  9. 9.
    S. K. Nayar, H. Murase, and S. A. Nene. Learning, positioning, and tracking visual appearance. In IEEE Int. Conf. on Robotics and Automation, San Diego, 1994.Google Scholar
  10. 10.
    M. Turk and A. Pentland. Eigenfaces for recognition. Journal of Cognitive Neuroscience, 3(1):71–86, 1991.Google Scholar
  11. 11.
    S. Yoshimura and T. Kanade. Fast template matching based on the normalized correlation by using multiresolution eigenimages. In Proceedings of IROS'94, pages 2086–2093, 1994Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Aleš Leonardis
    • 1
    • 2
  • Horst Bischof
    • 1
  1. 1.Institute for Automation Pattern Recognition and Image Processing GroupVienna University of TechnologyViennaAustria
  2. 2.Faculty of Computer and Information ScienceUniversity of LjubljanaLjubljanaSlovenia

Personalised recommendations