Image Identification Based on Color and Luminance Information by Using an Optical Correlator

A technique for distinguishing real-world images has been developed using a joint transform correlator. The real-world images, taken with a digital camera, are converted into luminance-intensity histograms and x-y chromaticity diagrams in the correlator system. These two-dimensional patterns are displayed on a spatial light modulator and are distinguished using the correlator system. The proposed identification techniques could distinguish an object image similar to the reference image of an apple among various fruits and vegetables, and could also distinguish tumor cell images from other cell images using the correlator system. These experimental results demonstrate the feasibility of using an optical correlator for the identification of complex real-world images containing large volumes of information.