All the ways that facial recognition can be fooled — and even hacked

  • India is working on the world’s largest facial recognition system that will comprise more than 10 million photos.
  • Once deployed, the system will automate the police procedures — match suspects’ photos with the database. But is the database secure enough to be trusted with a country’s security?
  • It is difficult to avoid a facial recognition system, However, over the past few years, researchers are looking out for new and innovative ways of doing so.

It is difficult to avoid a facial recognition system unless one keeps under a low light or stays home or hides somewhere. However, over the past few years, researchers are looking out for new and innovative ways of doing so.

This could pose a hurdle for countries which plan to use facial recognition as a part of public security system.

India is working on the world’s largest facial recognition system that will comprise more than 10 million photos. Once deployed, the system will automate the police procedures — match suspects’ photos with the database. But is the database secure enough to be trusted with a country’s security?
With new advancements in deep learning, facial recognition has become highly accurate. It is difficult to fool this technology, but not impossible. Here are tried and tested methods of evading it :



72153430

{{}}

View As: One Page Slides

​Hiding your face

​Hiding your face

Identifying partially hidden faces is no big deal for today’s technologies, so wearing glasses and scarves do not really help. But there’s really not much a computer can do if you go around hiding your face completely. If one goes around wearing a full face mask or uses these 3-D printed human face mask , you can hide from the cameras. But there’s a catch, it’s illegal to go around covering your face in most countries.


Credits: Neutrogena

​Hyperface, overwhelm by showing too many faces

​Hyperface, overwhelm by showing too many faces

You can trick the system by wearing something that has too many faces printed on it. The entrepreneurs of Hyperface project created clothes and accessories with too many fake faces on it. The use of numerous fake faces will make it difficult for the facial recognition system to recognise the real face.

​Tweaking your photos even before the database is launched

​Tweaking your photos even before the database is launched

The database can be fooled by subtly changing elements of a photo. A computer engineering student at the University of Toronto, Joey Bose, made an app that can modify certain elements of a photograph before it goes onto the Internet.

“The photos don’t look any different to the naked eye but the hidden features thwart detection systems,'' said Joey.

But there’s a catch. The technology works for only those people who haven’t uploaded their photos on social media sites like Facebook or Instagram. Many facial recognition systems use social media sites as a reliable source for database formation.

​Fighting tech with high tech glasses

​Fighting tech with high tech glasses

Normal glasses obviously don’t work.

A professor from Japan(Isao Echizen) invented a ‘privacy visor’ which reflected the light around the eyes when titled a little.

Reflectables, a company from Chicago also created glasses which blocks out the infrared light. But the product has not yet made it to the market.

Credits: Youtube/ikinamo

​Styling up with “Juggalos”

​Styling up with “Juggalos”

CV Dazzle is a company aiming to combine hair styles and heavy makeup to fool FR. Another style of makeup called ‘Juggalos’ using black paint on your chin which aims at fooling FR into misplacing the location of your jaw.

Even though there’s a chance that you may get around low tech FR which uses visible light, it’s no match for infrared FR which takes into consideration your entire face structure, primarily the contours, like in IPhone X. Heavy makeup hence, don’t work very well either.


Credits: Youtube/DarkArtboy

​Confusing the facial recognition by throwing extra infrared light

​Confusing the facial recognition by throwing extra infrared light


Since most facial recognition use infrared imaging to collect your facial features, it is possible to confuse it by throwing extra infrared light on the face which is invisible to the naked eye. That’s exactly what a group of researchers did in 2018. They embedded some infrared LEDs in a baseball cap which confused the computer vision and rendered the face unidentifiable in most cases.

Add Comment()
Comments ()
X
Sort By:
Be the first one to comment.
We have sent you a verification email. This comment will be published once verification is done.