Direct aspect-based 3-D object recognition
In this paper a method for 3-D object recognition based on Support Vector Machines (SVM) is proposed. Given a set of points which belong to either of two classes, a SVM finds the hyperplane that leaves the largest possible fraction of points of the same class on the same side, while maximizing the distance of the closest point. Recognition with SVMs does not require feature extraction and can be performed directly on images regarded as points of an N-dimensional object space. The potential of the proposed method is illustrated on a database of 7200 images of 100 different objects. The excellent recognition rates achieved in all the performed experiments indicate that the method is well-suited for aspect-based recognition.
Unable to display preview. Download preview PDF.
- 31.Bazaraa, M. and Shetty, C.M. Nonlinear programming (John Wiley, New York, 1979).Google Scholar
- 2.Brunelli, R. and Poggio, T. 1993. “Face Recognition: Features versus Templates,” IEEE Trans. on PAMI 15: 1042–1052.Google Scholar
- 3.Cortes, C. and Vapnik V. 1995. “Support Vector Network,” Machine learning 20: 1–25.Google Scholar
- 4.Edelman, S., Bulthoff, H., and Weinshall, D. 1989. “Stimulus Familiarity Determines Recognition Strategy for Novel 3-D Objects,” AI Memo No. 1138, AI Lab, MIT.Google Scholar
- 5.Huttenlocher, D.P., Klanderman, G.A., and Rucklidge, W.J. 1993. “Comparing Images Using the Hausdorff Distance,” IEEE Trans. on PAMI 15:850–863.Google Scholar
- 7.Osuna, E., Freund, R.,and Girosi, F. “Training Support Vector Machines: an Applications to Face Detection.” (Submitted to CVPR97).Google Scholar
- 10.Vapnik, V. The Nature of Statistical Learning Theory (Springer-Verlag, New York, 1995).Google Scholar