Researcher’s grant targets digital security

		Professor Jessica Fridrich's specialty is research in the field of digital security

A researcher whose work promises to spark major advances in the scope and security of digital communications, including digital audio, video and photography, will receive a $315,000 Air Force Office of Scientific Research grant.

The grant was awarded to Jessica Fridrich for her work on “Loss-Less Data Embedding and Crypto-Transformations for Steganography.” Fridrich is a research professor in the Department of Systems Science and Industrial Engineering in the Watson School of Engineering and Applied Science.

Steganography is the science of secret communication. Though some steganographic techniques, such as invisible inks and encryption, have been around for centuries, the first scientific papers on digital steganography and digital steganalysis — the investigation of hidden information — were published around 1993.

Fridrich, who earned her PhD in systems science at BU in 1995, is already a leader in the field. Her latest project is expected to perfect a new steganographic technique that will allow data to be embedded in a digital image and later extracted without altering the original image. All current data embedding techniques distort the image with “noise” and mean the inevitable loss of information during the extraction process.

Fridrich currently holds three patents, is applying for two more and has developed a Windows-based application, “Securestego,” that offers capabilities not available in other commercial steganographic software. Though she is still considering how best to market the package, several companies have already expressed interest in commercializing the software.

Incorporated in the Securestego application, which is regularly demonstrated at the Air Force’s Research Laboratory at Rome, New York, are a number of unique capabilities developed by Fridrich and her students. One of those applications, self-embedding, is a process that enables digital images that have been modified to revert to their unmodified state.

“If you have just used a software product to recolor a car from red to green thinking that you now have evidence supporting something that is not true, we can tell that,” Fridrich said.

“With self-embedded images, the original image can self-reconstruct, though the average user would have no way of knowing that it had survived their earlier alterations.”

Detection techniques 10 times better than those used by currently available products to ferret out hidden digital communications are also part of the package, she said.

These techniques are of particular interest to the military because they can help to protect against and intercept secret terrorist communications that rely on steganography.

Fridrich’s research, which is widely cited in her field, focuses on developing mathematical theory and appropriate equipment or apparatus to hide information in digital communications. She also works on ways to crack steganographic schemes of her own and others’ making.

“We need not only to build data-hiding schemes, but we also need to attack them to find out how secure the scheme is. This is, of course, very important to know in military applications, and also in civilian applications.”

By helping to ensure the authenticity of digital images and recordings through steganographic digital watermarking, Fridrich also expects her work to help make digital evidence more acceptable in courtrooms.

In a separate $99,200, project involving the Air Force and Kodak, Fridrich is designing a fragile watermark for digital cameras. Unlike robust digital watermarks, fragile watermarks “break” as soon as anything is done to alter the original signal or image. Cameras equipped with fragile watermarking capabilities will help to ensure the authenticity of digital photographs, lending them greater credibility in sensitive military and law enforcement applications, Fridrich said.