Marios Savvides: Biometrics Identification: An Effortless Future Security System

Marios Savvides: Biometrics Identification: An Effortless Future Security System


My name is Marios Savvides. I am a research professor in the Electrical
and Computer Engineering Department and CyLab here at Carnegie Mellon University. I’m also the director of the CyLab Biometrics
Center. What we do here at the lab is work on biometrics
identification. And biometrics is ways of how we can identify who you are based on physical
or behavioral characteristics. For example, fingerprint, face, iris, those
are all biometrics that can be used to identify who you are without you having to enter a
password. The trick, however, is how do we make biometrics
identification a really unconstrained, a thoughtless task? Well, there’s a negative stigma that Hollywood
has imposed on biometrics. And what we’re trying to do here is make biometric
identification a pleasant experience that is effortless. There’s been some technology that pushes biometric
identification one or two meters. What we’ve developed is a system that can
actually capture iris anywhere between 6 to 12 meters, and our system will actually pan
and tilt and capture the iris at that distance. So we’re the first system built that can do
that long range and capture iris images that are enrollment quality, identical to the kind
of image you capture at five inches away from the eye. One of our research thrusts that really excites
me, is trying to help law enforcement with some of the challenges they face day to day. If you ever watch the 8:00 news, you’ll see
many crimes that have happened that you know, they give out an image that’s grainy, that’s
low resolution, that’s off angle, these are all the kinds of challenges that law enforcement
face every day. And the current tools are just not able to
actually help find any leads about who the person may be. The tools that we’re developing now are actually
able to enhance the images are able to find possible matches given this low resolution
and off angle face. These tools can actually help find suspects
and hopefully close cases. Beyond law enforcement applications, our technology
has a wide audience anywhere from human robot interaction, to smart home surveillance. So if we can model all these factors, then
we’re making biometrics a very unconstrained and effortless technology that society is
going to embrace.

2 Replies to “Marios Savvides: Biometrics Identification: An Effortless Future Security System”

  1. Convenience store security cameras often can't capture enough biometric information to accurately identify a suspect. However, in our latest video, Marios Savvides, Professor of Electrical & Computer Engineering and Director of CyLab Biometrics Center, discusses the potential for enhancing criminal identification using long-range iris scanners. http://youtu.be/dpALtSd4r24

  2. I subscribed to this channel. Anyone looking for demos of Web Biometric Integrations can subscribe to my channel here https://www.youtube.com/channel/UCBeV1MXmOGBMPT-azuBAiYw?sub_confirmation=1

Leave a Reply

Your email address will not be published. Required fields are marked *