Nonlinear Preprocessing for Improved Filter Capacity of an Optical Correlator

Issue title: Optical Information Processing

Guest editors: S. KulakovGuest Editor

Article type: Research Article

Authors: Botha, Elizabeth C. | Bouma, Nicho G.

Affiliations: Department of Electrical and Electronic Engineering, University of Pretoria, Pretoria 0002, South Africa

Abstract: Optical correlation is an efficient parallel implementation of image classifiers. One formulation of the holographic filter in the optical correlator is the least squares discriminant function (LSDF), wherein the shape of the whole correlation function is specified when training the filter. An issue of particular interest is the capacity of the filter, or the number of training images that can be used while still retaining discrimination capability. To investigate the capacity of an LSDF filter, and to be more general than just testing on a particular type of image, the fractal dimension of the images was used as an objective complexity measure of the images that the filter is trained with. Nonlinear preprocessing, specifically edge detection, was used to increase the filter capacity and the discriminatory capabilities of the correlator. Simulation experiments of these issues are described and the results are discussed.

DOI: 10.3233/ICA-1996-3204

Journal: Integrated Computer-Aided Engineering, vol. 3, no. 2, pp. 117-128, 1996