FB pixel

The heart betrays even more in an age of biometrics. Lasers now ID people

Categories Biometric R&D  |  Biometrics News
The heart betrays even more in an age of biometrics. Lasers now ID people

Proving that, with the right technology, almost anything associated with a living being can used as a biometric signature, researchers are using effects of the heart to identify people.

In new research, civilian scientists say they can spot a deepfake by the way a face on video does — or does not — pulse with blood. And, separately, U.S. Department of Defense (DoD) researchers report they have developed two ways of identifying people by the unique way their hearts move clothing.

The first, potentially deepfake-spoiling, development, comes from a pair of Italian researchers who say they can detect a pulse in the face of someone caught on digital video.

Not being animated by a heart, digitally rendered human faces do not react to its rhythmic beat, according to Giulia Boato and Mattia Bonomi of the University of Trento.

Their own software is capable of spotting the lack of a pulse in videoed faces. The pair’s work has been published in the Journal of Electronic Imaging.

Deepfakes are digital video that blend the appearance of a person with a lifelike digital model of that person, using artificial intelligence and machine learning. The video can be manipulated convincingly to make it appear that even a very well-known person has done something they did not do.

As deepfakes grow in sophistication and number, worries grow also that soon no recorded evidence will be trustworthy. People can be added to scenes such as, most notoriously (so far), pornographic content, or removed from others — say a murder captured on security video.

This research follows related work, using an infrared laser to recognize an individual’s heartbeat, that was reported on last summer by MIT Technology Review.

Mention of the technology (named Jetson) dates back to a 2017 year-in-review publication released by the Defense Department’s office of combating terrorism technical support.

Military snipers depend on being able to identify, to the best of their ability, their military targets. They can use hard-won experience to identify people or, increasingly, they can use software and machine vision to recognize faces and gaits.

But glasses and a fake limp can defeat artificial intelligence pressed into this service. Harder to hide is the motion pattern of a person’s cardiac signature.

The Jetson device uses laser vibrometry. The technology has enabled researchers to positively identify people using an infrared laser, which is invisible, from 200 meters, The Economist reports, in an article mentioning Ideal Innovations, Inc. among vendors developing heartbeat biometrics. Vibrometry already is used to spot structural imperfections in turbines and warship canons.

In this case, the Jetson reads disturbances in worn fabric caused by a beating heart. Assuming a target’s heartbeat information is in a biometric library, a person can be positively identified 95 percent of the time under optimal conditions.

Shirts are light enough to betray the human heart; winter clothing is not. It is not clear if the laser spot, about the size of a quarter on a subject, must be placed only on the upper body. The spot must be placed on a person for about 30 seconds.

This is a novel use of technology that was first reported by USSR scientists in 1947. Clandestine Soviet teams used lasers played on windows to read vibrations in the glass. Conversations were easily recorded using this method until targeted nations began using pipes to direct music at glass to disguise voices.

Other current research, by Wenyao Xu at State University of New York at Buffalo, uses low-level Doppler radar to measure the unique dimensions of a person’s heart and its movement in the chest cavity from up to 30 meters away. Aluminum foil would prevent identification.

Related Posts

Article Topics

 |   |   |   |   |   |   |   | 

Latest Biometrics News


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Most Read From This Week

Featured Company

Biometrics Insight, Opinion

Biometrics White Papers

Biometrics Events

Explaining Biometrics