Abstract
The need to provide a neural network for pattern recognition with invariance to a new transformation, projectivity, is considered. This invariance is justified when working with object images that can appear rotated in relation to an axis contained in its own plane. An invariable relation to the transformation is found, the double ratio of four points, and incorporated to the network as a restriction to the weights. A projectivity invariant pattern classifier has been simulated. Besides, some considerations about high order neural networks are expounded.
Preview
Unable to display preview. Download preview PDF.
References
C.L. Giles and T. Maxwell,“Learning, Invariance, and Generalization in High-Order Neural Networks”. Applied Optics, vol. 26, No 23, pp. 4972–4978,(1987).
M.B. Reid, L. Spirkovska, and E. Ochoa,“Rapid Trainig of Higher-Order Neural Networks for Invariant Pattern Recognition”.Proc.Joint Int. Cnf. on Neural Networks,Washington D.C.,(1989), pp. 689–692.
S. Kollias, A. Tirakis and T. Milos, “An Efficient Approach To Invariant Recognition of Images Using Higher-Order Neural Networks”, Artificial Neural Networks. Elsevier Science Publishers B.V. (North-Holland) (1991), pp. 87–92.
P. Puig Adam,“Curso de Geometría Métrica II”.Ed. R. Puig,(1.965).
N.H. Goddard et Al,“Rochester Connectionist Simulator”, Technical Report 233, The University of Rochester (USA),(1989).
L. Spirkovska, M.B. Reid, “Coarse-Coding Applied to HONNs for PSRI Object Recognition”. Int.Joint Conf. on Neural Networks, Vol. II, p. A–931, Seattle, WA (1991).
X. Zhou, M.W. Koch, and M. W. Roberts, “A Selective Attention Neural Network for Invariant Recognition of Distorted Objects”, Int. Joint Conf. on Neural Networks, Vol. II, p. A–942, Seattle, WA 1991.
X. Zhou, N.W. Koch, and M.W. Roberts, “Selective Attention in High-Order Neural Networks for Invariant Object Recognition”, Int. Joint Conf. on Neural Networks, Vol. II, p. A–937, Seattle (WA), 1991.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1993 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Joya, G., Sandoval, F. (1993). Projectivity invariant pattern recognition with high-order neural networks. In: Mira, J., Cabestany, J., Prieto, A. (eds) New Trends in Neural Computation. IWANN 1993. Lecture Notes in Computer Science, vol 686. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-56798-4_195
Download citation
DOI: https://doi.org/10.1007/3-540-56798-4_195
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-56798-1
Online ISBN: 978-3-540-47741-9
eBook Packages: Springer Book Archive