The approach relies on an invisible watermark that is embedded into a digital file. Giuseppe Schirripa Spagnolo from the University of Roma Tre in Rome and his colleagues developed a system in which an image such as a company logo is first transformed into a computer generated hologram (CGH). A CGH is based on the light patterns made by a true hologram and a watermark manufactured in this way can be reconstructed in its entirety from only a fraction of the data. The researchers then encrypt the CGH data and add it to a digital image as a replacement to random, high frequency "noise" that doesn't affect image quality and is invisible to the human eye. Recipients of the digital image who know the key to the encryption can later reconstruct the watermark to ascertain whether the image has been tampered with.
In the images above, the original picture (top) was modified to change the colors of two cats (middle). After dividing the image into 16 quadrants, the researchers reconstructed the watermarks and determined which three quadrants (marked in yellow) had been corrupted. "We hope that this technique can be used to improve the reliability of photographs in the media" says co-author Lorenzo Cozzella. "Digital cameras could be developed so that an invisible watermark is added when a picture is taken. A newspaper buying a photo from a freelancer could then check for a watermark to confirm that it hasn't been tampered with to make it more newsworthy." The results appear in the latest issue of the Journal of Optics A.