Abstract

Wavelet transforms can be used to decompose a signal (image) into orthogonal components. When a small set of these components is used to form a wavelet transform sub-image, only specific characteristics of the image are retained. It is the identification of these characteristics that can be used to perform pattern recognition. A joint transform correlation system utilizing wavelet transforms is used to perform pattern recognition. An optical implementation of this system is described. Problems with respect to scale, rotation and translation are addressed and experimental results are presented.

© 1992 Optical Society of America