Menu

War Games

Capturing a shooter’s face in a firefight

facial recognition.jpg

A screenshot from the website of facial recognition company FaceFirst, which has teamed with Safety Dynamics to unveil a system to identify shooters.Airborne Biometrics Group

Fire a gun and your location can be pinpointed, your face photographed and your identity instantly determined -- all thanks to a new technology tag team.

When Safety Dynamics’ acoustic sensors detect a gunshot, they zero in on the shooter’s location and point a high-resolution camera at his or her face. Airborne Biometrics Group’s FaceFirst software then runs the shooter’s image against a biometric database to determine identity, even creating a new record if it can’t find one.

The companies say this is the world’s first such detector to instantly ID a shooter, and aim to make it available to law enforcement agencies and private physical security firms.

Gun-shot detection and facial recognition technologies have proliferated in recent years: Safety Dynamics competes with other solutions for threat recognition and localization such as Shotspotter, which relies on wide-area acoustic surveillance and GPS technology to triangulate the source of gunshots.

It goes a step further by adding facial recognition, an area in biometrics that has made huge progress in the past decade. Today’s systems often exploit 3D algorithms to boost accuracy, for example.

The company says the successful identification rate has improved from approximately 75 percent ten years ago to better than 95 percent by 2006 in government-sponsored tests. In tests run by the National Institute of Standards and Technology, its technology was given top ranking.

How it works
Ballistic signal sensors are positioned at a site, be it a public plaza or a urban neighborhood. When they pick up something, they relay the location’s azimuth and elevation in a split second to a camera.

An algorithm locates the face position and size for the camera, which snaps several images when possible. The camera aligns the image to center the eyes, rejecting its own picture and even sending an alert if they are too hard to read.

It can make other readability adjustments as well, improvements in light-dark contrast and distinguishing features, before extracting facial data such as nose length and mouth width. The FaceFirst system then beams the photo of the facial match and data about the shooter to a designated group -- hopefully speeding the perp’s apprehension.

For identifying the facial image in the private sector, company photographs or videos can form a reference database, or a company can use databases available through subscription.

FaceFirst can be entirely automated or include a human in the loop who can step into verify a facial recognition match when the software has not.

Who’s storing our pictures?

Facial recognition is used in airports and cities in the United States; it also infiltrates our daily lives through popular products from Facebook to iPhone and Android apps.

Indeed, Facebook recently suspended its dabbling in facial recognition in Europe in response to government reservations about the company’s biometric database on European citizens.

Last month, the U.S. Federal Trade Commission published a series of best practices recommendations for commercial companies developing facial image collections.

U.S. companies are discouraged from retaining client facial images indefinitely; the FTC cites the example of eyeglass websites that allow users to try on glasses.

Companies are also discouraged from identifying people where they would not want to be identified and give a bathroom or locker room as an example.

Software that identifies facial images such as an “app that allows users to identify strangers in public places, such as on the street or in a bar” is also discouraged by the FTC.

They are also encouraged to ensure the facial images are secure, notify people their faces are being captured and retained and obtain express permission for using the image.

Facial recognition technology has the potential to both better protect civilians and their privacy as well as invade it.

In this case, an alliance between gunshot detection tech and facial recognition is likely good news -- at least when it comes to catching criminals.

Ballet dancer turned defense specialist Allison Barrie has traveled around the world covering the military, terrorism, weapons advancements and life on the front line. You can reach her at wargames@foxnews.com or follow her on Twitter @Allison_Barrie.