Abstract
Selective attention shift can help neural networks learn invariance. We describe a method that can produce a network with invariance to changes in visual input caused by attention shifts. Training of the network is controlled by signals associated with attention shifting. A temporal perceptual stability constraint is used to drive the output of the network towards remaining constant across temporal sequences of attention shifts. We use a four-layer neural network model to perform the position-invariant extraction of local features and temporal integration of attention-shift invariant presentations of objects. We present results on both simulated data and real images, to demonstrate that our network can acquire position invariance across a sequence of attention shifts.
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
Bandera, C., Vico, F., Bravo, J., Harmon, M., Baird, L.: Residual Q-learning applied to visual attention. In: Proceedings of the 13th International Conference on Machine Learning, pp. 20–27 (1996)
Becker, S.: Implicit learning in 3D object recognition, The importance of temporal context. Neural Computation 11(2), 347–374 (1999)
Clark, J.J., O’Regan, J.K.: A Temporal-difference learning model for perceptual stability in color vision. In: Proceedings of 15th International Conference on Pattern Recognition, vol. 2, pp. 503–506 (2000)
Einhäuser, W., Kayser, C., König, P., Körding, K.P.: Learning the invariance properties of complex cells from their responses to natural stimuli. European Journal of Neuroscience 15, 475–486 (2002)
Földiák, P.: Learning invariance from transformation sequences. Neural Computation 3, 194–200 (1991)
Henderson, J.M., Williams, C.C., Castelhano, M.S., Falk, R.J.: Eye movements and picture processing during recognition. Perception and Psychophysics 65(5), 725–734 (2003)
Itti, L., Koch, C., Niebur, E.: A model of saliency-based visual attention for rapid scene analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence 20(11), 1254–1259 (1998)
Kikuchi, M., Fukushima, K.: Invariant pattern recognition with eye movement: A neural network model. Neurocomputing 38-40, 1359–1365 (2001)
Koch, C., Ullman, S.: Shifts in selective visual attention: Towards the underlying neural circuitry. Human Neurobiology 4, 219–227 (1985)
Körding, K.P., König, P.: Neurons with two sites of synaptic integration learn invariant representations. Neural Computation 13, 2823–2849 (2001)
Li, M., Clark, J.J.: Sensorimotor learning and the development of position invariance, poster presentation. In: The 2002 Neural Information and Coding Workshop, Les Houches, France (2002)
Minut, S., Mahadevan, S.: A reinforcement learning model of selective visual attention. In: AGENTS 2001, pp. 457–464 (2001)
Olshausen, B.A., Anderson, C.H., Van Essen, D.C.: A neurobiological model of visual attention and invariant pattern recognition based on dynamic routing of information. The Journal of Neuroscience 13(11), 4700–4719 (1993)
Olshausen, B.A., Field, D.J.: Sparse coding with an overcomplete basis set: A strategy employed by V1? Vision Research 37, 3311–3325 (1997)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2005 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Li, M., Clark, J.J. (2005). Learning of Position-Invariant Object Representation Across Attention Shifts. In: Paletta, L., Tsotsos, J.K., Rome, E., Humphreys, G. (eds) Attention and Performance in Computational Vision. WAPCV 2004. Lecture Notes in Computer Science, vol 3368. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-30572-9_5
Download citation
DOI: https://doi.org/10.1007/978-3-540-30572-9_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-24421-9
Online ISBN: 978-3-540-30572-9
eBook Packages: Computer ScienceComputer Science (R0)