Flag Manifolds for the Characterization of Geometric Structure in Large Data Sets
We propose a flag manifold representation as a framework for exposing geometric structure in a large data set. We illustrate the approach by building pose flags for pose identification in digital images of faces and action flags for action recognition in video sequences. These examples illustrate that the flag manifold has the potential to identify common features in noisy and complex datasets.
KeywordsTraining Sample Video Sequence Action Recognition Geodesic Distance Data Cloud
Unable to display preview. Download preview PDF.
- 2.J.R. Beveridge, B. Draper, J-M. Chang, M. Kirby, H. Kley, C. Peterson, Principal angles separate subject illumination spaces in YDB and CMU-PIE. IEEE TPAMI 29(2), 351–363 (2008)Google Scholar
- 4.J.-M. Chang, M. Kirby, H. Kley, C. Peterson, J.R. Beveridge, B. Draper, Recognition of digital images of the human face at ultra low resolution via illumination spaces, in Computer Vision – ACCV 2007, Tokyo, ed. by Y. Yagi, S.B. Kang, I.S. Kweon, H. Zha. Lecture Notes in Computer Science, vol. 4844 (Springer, 2007), pp. 733–743Google Scholar
- 7.I.M. James, The Topology of Stiefel Manifolds. London Math Society Lecture Note Series, vol. 24 (Cambridge University Press, Cambridge/New York, 1977)Google Scholar
- 8.F. Nater, H. Grabner, L. Van Gool, Exploiting simple hierarchies for unsupervised human behavior analysis, in IEEE CVPR, San Francisco, 2010, pp. 2014–2021Google Scholar
- 10.T. Sim, S. Baker, M. Bsat The CMU pose, illumination, and expression (PIE) database, in Proceedings of 5th International Conference on Automatic Face and Gesture Recognition, Washington, D.C., 2002Google Scholar
- 12.V.S. Varadarajan, Lie Groups, Lie Algebras, and their Representations. Graduate Texts in Mathematics, vol. 102 (Springer, New York, 1984)Google Scholar