Abstract
The sequential recognition method using the selective attention mechanism of Neocognitron (NC) is realizable and very effective for images of multiple targets. NC, however, is not applicable to general (large-scale, gray or color complex object) images, because of both a low reduction rate of coding scheme and an initial state dependency of competitive learning. In this paper, “hypercolumn model (HCM)” is introduced to modify NC and to apply the sequential recognition to general images. Two properties of the hierarchical self-organizing maps proposed by Lampinen: a coding scheme using topographic mapping and an initial state independence of learning resolve the two disadvantages of NC, respectively. Experimental results will show that a recognition accuracy of HCM for general images was from 10 percent to 30 percent higher than one of NC, and a sequential recognition performed well.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Bobick, A., Pentland, A. and Poggio, T.: “VSAM at the MIT Media Laboratory and CBCL: Learning and Understanding Action in Video Imagery”, Proc. Image Understanding Workshop, pp. 25–29, 1997.
Nayer, S. K. and Boult, T. T.: “OmnidirectionalVSAM Systems”, Proc. Image Understanding Workshop, pp. 51–61, 1997.
Carpenter, G. and Grossberg, S.: “A massively parallel architecture for a selforganizing neural pattern recognition machine”, Computer Vision, Graphics and Image Processing, Vol. 37, pp. 54–115, 1987.
Oja, E., Xu, L. and Kultanen, P.: “A new curve detection method: Randomized Hough Transform (RHT)”, Pattern Recognition Letters 11, pp. 331–338, 1990.
Oja, E., Xu, L. and Kultanen, P.: “Curve Detection by an Extended self-organizing Map and The Related RHT Method”, Proc. of Int. Neural Network Conf., Vol. 1, pp. 27–30, 1990.
Omori, T.: “Dual Representation of Image Recognition Process-Interaction of Neural Network and Symbolic Processing—”, Int. Symp. on Neural Information Processing NNET-14-1, pp. 50–53, 1992.
Tsuruta, N., Akagi, T., Taniguchi, R. and Amamiya, M.: “A Framework of Image Understanding Based on PDP Model—Proposition of ICE (Image CEntered) System—”, Int. Joint Conf. on Neural Networks, Vol. 1, pp. 177–180, 1993.
Kikuchi, M. and Fukushima, K.: “Neural Network Model of the Visual System: Binding Form and Motion”, Neural Networks, Vol. 9, No. 8, pp. 1417–1427, 1996.
Fukushima, K.: “Neocognitron: A self-organizing neural network Model for a mechanism of pattern recognition unaffected by shift in position”, Biological Cybernetics, Vol. 36, No. 4, pp. 193–202, 1980.
Fukushima, K.: “A neural network model for selective attention in visual pattern recognition”, Biological Cybernetics, Vol. 55, No. 1, pp. 5–15, 1986.
Lampinen, J. and Oja, E.: “Clustering Properties of Hierarchical Self-Organizing Maps”, Journal of Mathematical Imaging and Vision 2, pp. 261–272, 1992.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1999 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Tsuruta, N., Taniguchi, Ri., Amamiya, M. (1999). Hypercolumn model: A modified model of neocognitron using hierarchical self-organizing maps. In: Mira, J., Sánchez-Andrés, J.V. (eds) Foundations and Tools for Neural Modeling. IWANN 1999. Lecture Notes in Computer Science, vol 1606. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0098242
Download citation
DOI: https://doi.org/10.1007/BFb0098242
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-66069-9
Online ISBN: 978-3-540-48771-5
eBook Packages: Springer Book Archive