SIMPLE-SHAPE CLASSIFICATION BASED ON THE HUMAN VISUAL SYSTEM

This paper presents a new method suitable for shape classification, inspired by the early processing levels of the human visual system. It extracts a description for any simple 2-dimensional shape having a closed contour, regardless of its size, rotation and position, in affordable computational cost. The paper introduces a new approach to the modeling of the hypercolumns of the primary visual cortex, which requires significantly less computational burden and that is highly parallel. A new shape descriptor based on the relative angles of an object is also proposed. It produces close results for different shapes of the same object, it is proportion-flexible and it can identify distorted shapes correctly. Experimental results prove that the method is adequate for industrial production applications based on shape classification, as well as for shape-based image retrieval.

[1]  Ernesto Bribiesca,et al.  How to describe pure form and how to measure differences in shapes using shape numbers , 1980, Pattern Recognit..

[2]  Edmund T. Rolls,et al.  Position invariant recognition in the visual system with cluttered environments , 2000, Neural Networks.

[3]  Kunihiko Fukushima,et al.  Neocognitron: A hierarchical neural network capable of visual pattern recognition , 1988, Neural Networks.

[4]  Gert Cauwenberghs,et al.  Analog VLSI Cellular Implementation of the Boundary Contour System , 1998, NIPS.

[5]  Zia-ur Rahman,et al.  A multiscale retinex for bridging the gap between color images and the human observation of scenes , 1997, IEEE Trans. Image Process..

[6]  Hayaru Shouno,et al.  Training neocognitron to recognize handwritten digits in the real world , 1997, Proceedings of IEEE International Symposium on Parallel Algorithms Architecture Synthesis.

[7]  D. Hubel,et al.  Receptive fields and functional architecture of monkey striate cortex , 1968, The Journal of physiology.

[8]  Rolf P. Würtz,et al.  Corner detection in color images through a multiscale combination of end-stopped cortical cells , 2000, Image Vis. Comput..

[9]  Ernesto Bribiesca,et al.  A new chain code , 1999, Pattern Recognit..

[10]  Rodney M. Goodman,et al.  A real-time neural system for color constancy , 1991, IEEE Trans. Neural Networks.

[11]  Bartlett W. Mel.  SEEMORE: Combining Color, Shape, and Texture Histogramming in a Neurally Inspired Approach to Visual Object Recognition , 1997, Neural Computation.

[12]  Risto Miikkulainen,et al.  Contour integration and segmentation with self-organized lateral connections , 2004, Biological Cybernetics.

[13]  Stephen Grossberg,et al.  Synthetic aperture radar processing by a multiple scale neural system for boundary and surface representation , 1995, Neural Networks.

[14]  Laurent Itti,et al.  CINNIC, a new computational algorithm for the modeling of early visual contour integration in humans , 2003, Neurocomputing.

[15]  D. Hubel,et al.  Receptive fields, binocular interaction and functional architecture in the cat's visual cortex , 1962, The Journal of physiology.

[16]  D H HUBEL,et al.  RECEPTIVE FIELDS AND FUNCTIONAL ARCHITECTURE IN TWO NONSTRIATE VISUAL AREAS (18 AND 19) OF THE CAT. , 1965, Journal of neurophysiology.

[17]  Edmund T. Rolls,et al.  Invariant recognition of feature combinations in the visual system , 2002, Biological Cybernetics.

[18]  T. Poggio,et al.  Hierarchical models of object recognition in cortex , 1999, Nature Neuroscience.

[19]  Laurent Itti,et al.  A Model of Contour Integration in Early Visual Cortex , 2002, Biologically Motivated Computer Vision.