Share Email Print
cover

Proceedings Paper

Unsupervised learning of a steerable basis for invariant image representations
Author(s): Matthias Bethge; Sebastian Gerwinn; Jakob H. Macke
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

There are two aspects to unsupervised learning of invariant representations of images: First, we can reduce the dimensionality of the representation by finding an optimal trade-off between temporal stability and informativeness. We show that the answer to this optimization problem is generally not unique so that there is still considerable freedom in choosing a suitable basis. Which of the many optimal representations should be selected? Here, we focus on this second aspect, and seek to find representations that are invariant under geometrical transformations occuring in sequences of natural images. We utilize ideas of 'steerability' and Lie groups, which have been developed in the context of filter design. In particular, we show how an anti-symmetric version of canonical correlation analysis can be used to learn a full-rank image basis which is steerable with respect to rotations. We provide a geometric interpretation of this algorithm by showing that it finds the two-dimensional eigensubspaces of the average bivector. For data which exhibits a variety of transformations, we develop a bivector clustering algorithm, which we use to learn a basis of generalized quadrature pairs (i.e. 'complex cells') from sequences of natural images.

Paper Details

Date Published: 12 February 2007
PDF: 12 pages
Proc. SPIE 6492, Human Vision and Electronic Imaging XII, 64920C (12 February 2007); doi: 10.1117/12.711119
Show Author Affiliations
Matthias Bethge, Max Planck Institute for Biological Cybernetics (Germany)
Sebastian Gerwinn, Max Planck Institute for Biological Cybernetics (Germany)
Jakob H. Macke, Max Planck Institute for Biological Cybernetics (Germany)


Published in SPIE Proceedings Vol. 6492:
Human Vision and Electronic Imaging XII
Bernice E. Rogowitz; Thrasyvoulos N. Pappas; Scott J. Daly, Editor(s)

© SPIE. Terms of Use
Back to Top