Open Competition 3 - Information Technology

3-D Face Recognition for Security Screening

Sponsor: Advanced Digital Imaging Research, LLC

• Project Performance Period: 6/1/2004 - 5/31/2007
• Total project (est.): $3,189,536.00
• Requested ATP funds: $1,981,989.00

Automatic facial recognition systems, a potential tool for security screening at airports and elsewhere, often fail in practical use because lighting, expression, pose, and other conditions are variable. Moreover, existing databases of the faces of wanted criminals and terrorists are inadequate. Advanced Digital Imaging Research plans a three-year project to develop and test a real-time, three-dimensional (3D) face recognition system camera that improves performance by combining the advantages of both 3D and 2D imaging for unobtrusive use at airport security checkpoints. A high-resolution spectroscopic camera will be used to obtain 3D images (or textured surface maps) of faces. Images will be processed to eliminate accessories and facial hair, and quantitative measurements will be derived for comparison against 3D databases. In addition, to remedy the lack of sufficient 3-D data on wanted criminals, existing 2D images (e.g., mugshot and profile) will be combined to make 3D images, and 3D images will be projected onto 2D images to make "synthetic mugshots" for comparison to 2D databases. Technical challenges are involved primarily in the integration of the system, which requires image acquisition hardware, 2-D and 3-D databases, and face recognition software, all packaged with a convenient user interface. The system will need to achieve high accuracy in real time, with low false alarm rates. The University of Houston will be subcontracted to develop new face recognition algorithms. Prototypes will be tested at Ellington Field Airport and Hobby Airport in Texas. ATP support is needed because funding is not available from ADIR's parent company or other private sources. Improved screening would enhance the safety of air travel and reduce airport delays, which cost airlines, passengers, and shippers $5 billion in 1998. If successfully developed, the technology also could be used at seaports, nuclear facilities, commercial buildings, and military installations, as well as at schools and public events.

For project information:

Kenneth R. Castleman, (281) 535-1889
castleman@adires.com

ATP Project Manager

Carlos Grinspon, (301) 975-4448
carlos.grinspon@nist.gov