The US Military Is Building Long-Range Facial Recognition That Works In The Dark

Navin Bondade Artificial Intelligence, Data Science, Deep Learning, Machine Learning

The United States military is building a super-charged facial recognition system that could detect the faces of individuals in a long-range and in the dark conditions, the system will be ready for action as soon as next year.

According to OneZero, the U.S. military is spending more than $4.5 million to develop a facial recognition system that reads the pattern of heat being emitted by faces in order to identify people in trivial conditions.

In order to build this system, the US Military has given one grant $2.8 million to Cyan Systems, a California-based company specializing in infrared sensor technology. The other gives regular government contractor Polaris Sensor Technologies $1.5 million to develop the facial recognition system.

Polaris develops electro-optic sensors & novel imaging sensors using polarization to offer image recognition. Cyan Systems is focused on infrared sensor technology and hardware for high-definition imaging.

The start date for both contracts is listed as September 2019, with an estimated end date of September 2021. According to the details of the request, the DFBA is directly overseeing work on the technology.

Last year, OneZero reported that the DFBA is responsible for the entirety of DoD’s facial recognition, fingerprint, and DNA analysis efforts. Will Graves, listed as the primary contact in the Army request for the technology, has represented DFBA at a number of industry conferences since 2018.

The facial recognition system work in the dark and it analyzes infrared images of a person’s face to see if they’re a match for anyone on a government watchlist, such as a known terrorist.

Not only will the finished system reportedly work in the dark, through car windshields, and even in less-than-clear weather conditions — but it’ll also be able to ID individuals from up to 500 meters away.

The “sensors should be demonstrable in environments such as targets seen through automotive windshield glass, targets that are backlit, and targets that are obscured due to light weather (e.g., fog).”

The range of detection needs to span 10 to 500 meters.

The Department of Defense is expecting a device small enough to be carried, but capable of reading the heat pattern of a face and allowing it to be identified. It’s expected this will be achieved through a combination of infrared imaging and thermal image-processing techniques.

Infrared, long-range facial recognition has the potential to drastically increase the military’s ability to identify people who pass even within a quarter of a mile of military personnel. According to the DoD request, the new device will be used to identify those on a watch list rather than combing through the DFBA’s entire database.

“Fusion of an established identity and information we know about allows us to decide and act with greater focus, and if needed, lethality,” the DFBA’s director wrote in presentation notes last year.