MSU researchers develop whole-body biometric recognition system with $12M grant

Michigan State University researchers have developed technology for long-range biometric identification, supported by a four-year federal grant from the Intelligence Advanced Research Projects Activity (IARPA) worth $12 million under its Biometric Recognition and Identification at Altitude and Range (BRIAR) program, according to an announcement.
The goal of the program is to develop end-to-end software systems that can detect individuals and extract “biometric signatures” from the entire body, such as gait and body shape, as well as the face for multimodal biometric matching.
The system incorporates the use of drones for recognizing individuals at a distance. The elevation allows for increased altitude and range for imaging, but the vantage point mainly captures the tops of people’s heads, limiting the capacity to identify people through facial recognition alone.
By combining the modalities of face, body shape, and gait, the system can better identify people from the elevated vantage point. The system, called FarSight, accepts input from drone-captured videos and generates a candidate list of identities from a database.
“We were surprised to learn that a person’s height, build and gait often contain more useful biometric information than their face when capturing images from a long distance,” said Xiaoming Liu, an MSU Research Foundation professor leading the project. Prominent biometrics researchers Anil Jain and Arun Ross are also part of the team.
The system accounts for low-quality imagery, turbulence, varying angles and distances, and limited operational data in its identification processes.
In the project’s current phase, the system can “look at one person from about 300 to 400 meters away,” while in the next phase, “we’ll be looking at multiple people together at a range of 600 to 700 meters,” says Liu.
Potential use cases include helping law enforcement agencies find missing person(s) or fugitives.
The collaboration on biometric surveillance technologies by government agencies and academic institutions drew criticism from Congress earlier this year.
Article Topics
biometric identification | drones | facial recognition | gait recognition | IARPA | long-distance | MSU | research and development
Comments