A Novel Software for Detection of Hidden Messages Within Digital Images

Steganography is the art and science of hiding information for the purpose of covert communication. Steganography has a long history that includes techniques such as invisible inks, null ciphers, microdots, and embedded messages. With the advent of the digital age, steganography has seen a resurgence, with applications in the protection of copyrights in digital media, unobtrusive military and intelligence communication, covert criminal communication, and for the trafficking of illegal pornography. Modern day steganographers typically embed hidden messages within digital images. Composed of millions of pixels, a digital image provides an ideal cover for hiding secret messages.



Dartmouth investigators have developed a computer program that is capable of detecting the presence of hidden messages embedded within digital images. This program consists of three basic stages. In the first stage higher-order statistical models of natural images are constructed. In the second stage, a classifier is automatically constructed to “optimally'' differentiate between the statistics of images with and without hidden messages. In the third stage,the same statistics are extracted from an image of interest, and fed to the classifier. The classifier outputs the probability of the image containing a hidden message. The novelty of our work is in the specific statistical models that we employ, and in our ability to detect hidden messages independent of the specific image format or hiding technique.



We have tested our program on a set of 1,800 images and four different hiding algorithms. Depending on the size of the message relative to the cover medium, we are able to detect the presence of a hidden message with an average accuracy of 65% and a less than 0.001% false positive rate. There is no inherent reason why these techniques can not be generalized to digital audio or video.

Dartmouth College is looking for an industrial partner interested in licensing this technology. (Ref: J197

Type of Offer: Licensing



Next Patent »
« More Software Patents

Share on      


CrowdSell Your Patent