Advanced Technology Program ATP Home Page NIST Home Page

Project Brief


Open Competition 2 - Information Technology

Secure, Robust, Forensic Watermarking


Develop watermarking identification technology for superimposing invisible data patterns on image sequences -- identifying the source of each specific image sequence copy -- with sufficient fidelity, security, and robustness to be applied to digital films to track and deter theft.

Sponsor: Sarnoff Corporation

201 Washington Road
CN 5300
Princeton, NJ 08543-5300
  • Project Performance Period: 10/1/2001 - 3/31/2003
  • Total project (est.): $1,489,124.00
  • Requested ATP funds: $861,703.00

Piracy of copyrighted material is a major impediment to the digital distribution of movies in theaters and other channels such as Internet video. Digital content can be protected by encryption only up until the point of presentation, where it becomes vulnerable to unauthorized copying. To deter piracy at this point, Sarnoff Corp. plans to develop a watermarking identification algorithm that will superimpose invisible data patterns on image sequences as a movie is being shown to provide means of determining the date, time, and place of post-decryption theft, such as by camcorders used in theaters. Existing watermarking technologies are not robust against severe distortion, such as that from camcorder capture, and cannot carry sufficient data to identify points of theft. In a one-year project, Sarnoff plans to overcome these limitations by embedding the data into the low spatial and temporal frequencies of the imagery, a novel approach. To ensure that the watermark is imperceptibly embedded, the company will use its Emmy-award winning vision modeling capability developed in a previous ATP project to predict the visibility of the data patterns. A number of viable watermarking patterns will be identified so that "frequency hopping" can be used to prevent jamming. Measurements, modeling, and other analysis will be performed for each pattern. ATP funding is needed to demonstrate and prove the technology before movie companies will be willing to invest in its further development. Professor Bede Liu of Princeton University, a leading expert on watermarking security, will be subcontracted to ensure that the algorithm is compatible with the security requirements of digital video distribution. If successfully developed and commercialized, the new technology will reduce movie piracy losses, estimated at $3 billion annually and growing, and accelerate the adoption of digital cinema, enabling substantial savings in film printing costs and spurring new revenues for the movie industry, a substantial contributor to the U.S. economy. The new technology also could be useful in other image-marking applications, such as in the security and health-care industries, and improve the secure digital distribution infrastructure for e-commerce.

For project information:
Thomas Lento, (609) 734-3178
tlento@sarnoff.com

ATP Project Manager
Barbara Cuthill, (301) 975-3273
barbara.cuthill@nist.gov


ATP website comments: webmaster-atp@nist.gov
Privacy Statement / Security Notice NIST Disclaimer NIST Information Quality Standards
NIST is an agency of the U.S. Commerce Department