FIRST AIRED: March 25, 2018

Nice work! Enjoy the show!

×

You’re busy. We get it.

Stay on top of the news with our Editor’s Picks newsletter.

US Edition
Intl. Edition
Unsubscribe at any time. One click, it’s gone.

Thanks for signing up!

×

Transcript

00:00:00
>> Computers that can spot a weapon without human supervision. That's what these researchers at Spain's University of Grenada say they've devised, aimed at stopping an attack before it's happened. In this mockup, every time a gun or knife appears in vision, a colored box highlights it on screen and sends an alarm to those monitoring.
00:00:18
Professor Francisco Herrera is on the team.>> For example, it may be possible to use in airports to detect and let's suppose we have a camera system in the underground of the airport and people go to the airport with a car and they take some arms from the car.
00:00:34
It is possible to send immediately an alarm because one person is passing in their hand a knife.>> Object recognition technology itself isn't new. But according to its creators, this system is self-training. By using an artificial intelligence deep learning model, it's able to improve itself and learn from its successes.
00:00:57
The team initially tested their new program on low-quality recordings of Hollywood blockbusters like Pulp Fiction and Mission Impossible, and it still spotted someone holding a gun 96% of the time. They now plan to test their system in crowded areas, and several surveillance companies have already expressed their interest in using it.