----- Forwarded message from Bob Margolis <rttyman@wwa.com> ----- Date: Mon, 25 May 1998 12:54:47 -0500 From: Bob Margolis <rttyman@wwa.com> To: spooks@qth.net Subject: [Spooks] Digital Steganography Secret Messages: Army Researchers Discover Advance Technique To Relay Information Army researchers bring new meaning to the phrase, 'a picture's worth a thousand words.' An advanced technique for hiding secret messages inside of images is a new discovery made recently by U.S.Army Research Laboratory (ARL) researchers and a Federated Laboratory partner from the University of Delaware. "As a method to provide secure communications on the battlefield, we are pursuing research in the area of digital steganography," says Lisa Marvel, an electronics engineer at ARL's Information Science & Technology Directorate. The Greek word, steganography, means covered writing. "Our initial system involves embedding hidden messages within images. Images that are typically transmitted over the battlefield, such as weather or terrain maps," she says. The hidden message, be it an authentication mark, time stamp or new orders to a commander, is completely undetectable and resistant to removal and arbitrary decoding. Only the intended recipient holding the 'key' can decode the hidden message. Consider for a moment the impact of this new capability. U.S. Forces are pent behind enemy lines and out numbered three to one. The commander is not able to communicate their situation to base, because an enemy listener is detected on the communication network. The commander orders a message be sent through the Stego Communication System. In less than 45 seconds a soldier types the message on a laptop computer, hides the message in a map and sends it to base undetected. U.S. troops are deployed with enemy forces unaware of the impending ambush. U.S. soldiers are recovered with no casualties. Steganography is part of ARL's defensive information warfare efforts under the Federated Laboratory Program. The research began in February 1997 and is projected to continue over the next two years. "The result of this basic research can be packaged as software that can run on any laptop, anywhere. The main focus for the next two years is to increase the amount of undetectable information that we are able to hide in an image and develop techniques that make messages even more resilient to any kind of corruption," Marvel says. Typically, basic research is projected for use five years out. But, Marvel says this particular work is implementable in a shorter timeframe. She predicts that when the research is completed in two years it will just be a matter of software rewrite and packaging, which will take about 6 months before it can actually become a true product. Marvel credits the collaborative work, commissioned by the Federated Laboratory, for the successful progress of the project. Marvel developed the initial test system, guided by Dr. Charles Boncelet, Professor of electrical engineering at the University of Delaware. "Dr. Boncelet brought in some specific insights and his expertise in random signals and noise truly contributed to the success of the work," Marvel says. Dr. Charles Retter another ARL engineer provided his expertise in low rate error correcting codes, "which is key to how we are able to extract the message from the image without having the original image," she adds. Digital steganography is a new topic of research that incorporates ideas from communications, cryptography, information theory, and signal processing. So far, very little information exists in the formal literature on this topic, unlike digital watermarking which is currently a very active area of research. Although similar, steganography differs from watermarking in several aspects. For example, digital watermarking is used to embed a visible or invisible signature into a medium in order to indicate ownership, such as the station insignia seen in the bottom corner of a television screen during programming. Steganography, on the other hand, is used to encode a hidden message typically much larger than a signature, Marvel explains. In preparing to file for a patent, the ARL researchers are conducting the required background research on the prior art. "We found that a few basic techniques used over the years work pretty well, but are not nearly as sophisticated or complete. It appears that our technique is unique," Marvel concludes. --- Submissions spooks@qth.net ----- End forwarded message ----- -- Dave Emery N1PRE, die@die.com DIE Consulting, Weston, Mass. PGP fingerprint = 2047/4D7B08D1 DE 6E E1 CC 1F 1D 96 E2 5D 27 BD B0 24 88 C3 18