Biologically Inspired Bayes Learning and Its Dependence on the Distribution of the Receptive Fields
In this work we explore the dependence of the Bayesian Integrate And Shift (BIAS) learning algorithm on various parameters associated with designing the retina-like distribution of the receptive fields. The parameters that we consider are: the rate of increase of the sizes of the receptive fields, the overlap among the receptive fields, the size of the central receptive field, and the number of directions along which the centers of the receptive fields are placed. We show that the learning algorithm is very robust to changes in parameter values and that the recognition rates are higher when using a retina-like distribution of receptive fields compared to uniform distributions.
KeywordsReceptive Field Recognition Rate Directional Distribution Circular Distribution Foveal Vision
Unable to display preview. Download preview PDF.
- 2.Daugman, J.G.: Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimensional visual cortical filters. J. Optical Soc. Am. 2(7), 1,160–1,169 (1985)Google Scholar
- 3.Gomes, H.M.: Model learning in iconic vision. Ph. D. Thesis, University of Edinburgh, Scotland (2002)Google Scholar
- 4.Land, M.F., Nilsson, D.-E.: Animal Eyes. Oxford University Press, Oxford (2002)Google Scholar
- 5.Lee, T.S.: Image representation using 2d gabor wavelets. PAMI 18(10), 1–13 (1996)Google Scholar
- 6.Lindeberg, T., Florack, L.: Foveal scale-space and the linear increase of receptive field size as a function of eccentricity. Tech. Report, Royal Institute of Technology, S-100 44 Stockholm, Sweden (1994)Google Scholar
- 8.Neskovic, P., Wu, L., Cooper, L.: Learning by integrating information within and across fixations. In: Proc. ICANN 2006 (2006)Google Scholar
- 10.Sekuler, R., Blake, R.: Perception. McGraw-Hill Companies Inc., New York (2002)Google Scholar
- 11.Serre, T., Wolf, L., Poggio, T.: Object recognition with features inspired by visual cortex. In: Proc. CVPR 2005 (2005)Google Scholar