In the approach developed by Prof. Joseph Rosen of BGU's Department of Electrical and Computer Engineering and Prof. Bahram Javidi of the University of Connecticut, a block of encrypted information, such as the bearer's fingerprint pattern or other unique personal facts, can be concealed within a picture on a document. Because this information is relevant to the bearer alone, use of the ID by a person resembling the card holder is easily unmasked. Moreover, since the complex keys used to encrypt and decrypt the hidden information are known only to the issuer, it would be next-to-impossible to forge an ID that would pass through the system.

According to Rosen, the new development combines of two well-known methods of representing data. One is the halftone image, a 2-dimensional (2D) pattern of larger and smaller dots used in reproducing pictures. The other is a 2D barcode, in which a checkerboard of tiny dots and spaces represents digital information.

Barcodes are familiar to all of us: they are the long string of lines printed on packaging labels for product identification. But a 2D barcode comprised of small dots can record much more information than the string of lines and is an inexpensive way to provide extensive data about a person or a manufactured product.

In this new development, Rosen and Javidi have combined the 2D halftone image and barcode by slightly shifting the positions of the arrangement halftone dots.

The concealed barcode information can be retrieved using what is known as a 2-D spatial correlator, which contains a confidential filter function that deciphers the concealed image.

The new technology, Rosen claims, is very robust as even a damaged ID picture or a half-covered picture contains sufficient hidden data to retrieve the encrypted information, This is because the entirety of secret data is distributed throughout the picture. However, if the original picture is not whole, the hidden picture will be reproduced with lower quality. Another advantage of this development is that both optical and computational approaches can be used to reveal the hidden data.

A patent has been submitted for approach to conceal an image within an image. The advance has been reported in Applied Optics 40, 3346-3353 (2001), Journal of the Optical Society of America.

Published by Israel's Business Arena on 10 December 2001