Orientation Competition in Cortical Filters - An Application to Face Recognition.pdf

Orientation Competition in Cortical Filters -An Application to Face RecognitionN. Petkov, P. Kruizinga, T. Lourens Department of Mathematics and Computer ScienceRijksuniversiteit GroningenBlauwborgje 3P.O. Box 800, 9700 AV GroningenThe Netherlands In: Computing Science in the Netherlands 1993, Nov. 9-10, 1993, Utrecht (Stichting MathematischCentrum: Amsterdam, 1993) pp. 285-296. 0 Orientation Competition in Cortical Filters -An Application to Face RecognitionN. Petkov, P. Kruizinga, T. LourensDepartment of Mathematics and Computer ScienceRijksuniversiteit GroningenP.O. Box 800, 9700 AV GroningenThe NetherlandsAbstractA biologically motivated, computationally intensive approach to computer vision is developedand applied to the problem of automatic face recognition. The approach is based on the use of two-dimensional Gabor functions which model the receptive eld functions of simple cells in the primaryvisual cortex of mammals. The convolutions of an input image with a set of antisymmetric visualreceptive eld functions (imaginary parts of Gabor functions) become the subject of thresholding andorientation competition. The developed cortical lters deliver highly structured information which isused for ecient feature extraction and representation in a lower dimension space. Applied to facerecognition, the method gives a recognition rate of 98.5% on a large database of face images.1 IntroductionLarge scale computer simulations are nowadays a well established research tool in natural and engineeringsciences such as physics, chemistry, astronomy, uid dynamics, electrical engineering, etc. In the 1990s,proclaimed to be the decade of the brain, the insights in the microstructure of the brain provided byneurophysiological and neurobiological research together with the progress in mathematical models ofarticial neural networks may open new opportunities for computational science. In the years to comelarge scale computer simulations may become an instrument of neuroscience,