Computer Vision and Biometrics

Terrorists and criminals depend on anonymity to do their destructive work. Fortunately, the Department of Defense (DoD) and Department of Homeland Security (DHS) are uncloaking offenders using Booz Allen’s artificial intelligence (AI). We support the mission of citizen protection through advanced human identification.

The Booz Allen team has been helping DoD identify terrorists since the early 2000s. “Now we have a 24/7 defense operations center, staffed by scientists who have 15 years of experience developing advanced identification algorithms for DoD,” says Jonathan Levitt, chief scientist. “When events happen, we get requests in the middle of the night and can return analysis reports within hours.”

The team includes scientists, subject matter experts, analysts, and developers, many with PhDs in fields like computer vision and machine learning.

“Besides conducting intelligence analysis, we're constantly developing new tools and capabilities,” Jonathan says. The team builds next-generation biometrics and forensics solutions, including sensors, algorithms, and imaging capabilities. They also develop custom biometric systems and solutions for defense, intelligence, and law enforcement. 

The team’s work to identify terrorists was so successful that DHS asked them to help identify perpetrators of child exploitation. The two missions have similar requirements: Fast turnaround, ability to analyze mixed media, and techniques that work even on unconstrained data—meaning we’re not in control of how the subject was recorded. “If you take a standard photo, you place the person a couple of feet away, facing the camera. In contrast, our media source may be of poor quality: Low resolution, bad lighting, and a subject who isn’t showing their face,” explains Jonathan.

Although the DHS program launched only recently, it already provided a measurable improvement in identification rates for DHS. “We keep moving the field forward,” says Jonathan. “Thanks to our work across various client missions, we have success in analyzing media that was so low quality it was once thought to have no value.”

When asked about the most rewarding part of his job, he is quick to reply. “Helping prevent events that shouldn’t happen,” he says. “We’ve been able to stop really bad people from doing really bad things.”