Image feature extraction with various wavelet functions in a photorefractive joint transform correlator

The wavelet transform has found a lot of uses in the field of optics. We pr esent an experimental realization of employing Variant wavelet filters into the object space of a photorefractive joint transform correlator to realiz e image feature extraction. The Haar's wavelet, Roberts gradient and Mexica n-hat wavelet are employed in the experiment. Because of its good optical p roperties, the photorefractive crystal Bi12SiO20 is used as the dynamic hol ographic medium in the Fourier plane. Both scene and reference have been de tour-phase encoded in a liquid crystal television in the input plane. Compu ter simulations, experimental results and analysis are presented. (C) 2000 Elsevier Science B.V. All rights reserved.