A credit card verifier structure using diffraction and spectroscopy concepts

Abstract

Listen

Translate article

We propose and experimentally demonstrate an angle-multiplexing based optical structure for verifying a credit card. Our key idea comes from the fact that the fine detail of the embossed hologram stamped on the credit card is hard to duplicate and therefore its key color features can be used for distinguishing between the real and counterfeit ones. As the embossed hologram is a diffractive optical element, we choose to shine one at a time a number of broadband lightsources, each at different incident angle, on the embossed hologram of the credit card in such a way that different color spectra per incident angle beam is diffracted and separated in space. In this way, the number of pixels of each color plane is investigated. Then we apply a feed forward back propagation neural network configuration to separate the counterfeit credit card from the real one. Our experimental demonstration using two off-the-shelf broadband white light emitting diodes, one digital camera, a 3-layer neural network, and a notebook computer can identify all 69 counterfeit credit cards from eight real credit cards.