Facial recognition technology can trace its roots as far back as the 1960s. The ability to implement it large scale and ubiquitously only really came into being in the modern computer age, of course, and in the present it seems to be nearly everywhere, with fears that it will become even more widespread in the future. But many of us are unaware of just what is already being done with the technology in terms of both strengths and limitations.
10. We Have Facial Recognition For Cats
Is any technology even worth your time if it can’t be used for the benefit of cats and dogs? Facial recognition luckily fits the bill here as the technology has been implemented towards the end of reuniting lost pets with owners.
The concept is as simple as you might expect. There’s an app which allows you to upload photos of a lost pet. Once the pics are in the database if someone finds your pet, the animal can be matched with facial recognition to your fur friend and you can be reunited.
That’s not the only use, either. There’s a cat feeder that makes use of facial recognition to monitor the eating habits of your pets. This could range from alerting you that one cat isn’t eating enough to preventing another cat from stealing everyone’s food by shutting off the supply when it recognizes the same hungry boy coming back for seconds.
9. The FBI Spent $1B on Facial Recognition That Could Only Match Photos to Photos
Law enforcement obviously has a vested interest in facial recognition and while there is a ton to be said about the various legal implications, privacy concerns, nefarious prospects and more of that whole can of worms, there’s also something more basic and fundamental to discuss and that’s just how well it works in the first place.
Now, in the present, the technology has definitely improved and there are examples of it performing very well. But if we go back to when the FBI first invested in it, things get a little uglier. The Bureau spent about $1 billion on software that couldn’t actually recognize faces unless they were matched to quality photos.
You needed to take one good quality, front-facing, clear photo of a person and the computer could match it to another photo that met the same requirements. You might recognize this as something a human could also do. The computer’s only advantage might be the ability to do it quickly.
If the computer doesn’t have a second, quality photo to match it to, then there is no match. This was discovered in Boston after the Boston Marathon bombing in 2013.
The FBI had access to photos of the two suspects and they were there, caught on video, but the facial recognition software couldn’t match them because the surveillance footage wasn’t high quality, thus making it most useless.
According to one FBI source, the billion dollar software was only good for matching “driver’s license and passport photos to other driver’s license and passport photos.”
The FBI had access to photos of the two suspects and they were there, caught on video, but the facial recognition software couldn’t match them because the surveillance footage wasn’t high quality, thus making it most useless at the time. According to one FBI source, the billion dollar software was only good for matching “driver’s license and passport photos to other driver’s license and passport photos.”
8. Juggalo Makeup Defies Some Kinds of Facial Recognition Technology
If you’re looking to trick facial recognition for some kind of lawful and non-nefarious reasons, look into becoming a fan of the Insane Clown Posse. The makeup worn by the band and their fans, who you may know by the more colorful name of “Juggalos” has proven to be something facial recognition has a tough time with.
Facial recognition relies on contrast for a lot of the ways it differentiates faces to identify them. Juggalo makeup uses a white base with black bands and other shapes to create a sort of monochromatic clown motif. Because it creates stark but fake contrast, facial recognition can’t tell who it’s looking at.
This revelation came to light when LiveNation and Ticketmaster bought military facial recognition to use at concert entry points. Juggalo makeup, especially contrasting black and white around the mouth and chin, confuses facial recognition by obscuring your jawline which is one of the major areas the technology focuses on.
Of course, there are kinds of technology that work differently and don’t focus on contrast which would render Juggalo makeup useless, but at least you’d still be visibly supporting the Insane Clown Posse.
7. Eyebrows Are Vital to Facial Recognition
For some people, having nice eyebrows is a big deal. For others, having fuzzy caterpillars shading their eyes is just fine. For facial recognition, eyebrows are of the utmost importance. While eyebrows may not seem like they would play a key role in facial recognition when compared to things like eyes, noses, lips and overall face shape, that’s not the case.
Biologically speaking, eyebrows are vital for identifying faces. The presence of hair, otherwise missing from most human faces, provides a major point of contrast to someone looking at you. Even person to person, eyebrows are key for understanding another face. It makes sense that software also takes this into account.
While eyebrows haven’t always been used as a key identifier in various softwares, the Covid-19 pandemic made some researchers tweak the way their software worked since many people were wearing masks which eliminated things like mouths, jawlines and noses as identifying features.
6. Taylor Swift Used Facial Recognition to Identify Stalkers at Concerts
We already saw that concerts use facial recognition, and it’s more than just for Juggalos. Taylopr Swift, one of the biggest artists in the world right now, has been using it to identify potential stalkers.
Back in 2018, Swift had kiosks set up outside of her performance at the Rose Bowl. As people filtered in, the kiosks were showing off clips of her dance routines but they were also secretly scanning faces. Facial recognition was used to cross reference all the faces against a list of known stalkers she’s had to deal with in the past so security could be ready.
The same technology is being used at NASCAR events, in malls and at sporting events for a variety of very non-specific purposes that include both security and also advertising. So if you have gone out to a public event in the last few years there’s a good chance that some company you have never heard of has your name and photo on file along with a list of things it has determined you like. By 2019 the company had already collected data on 110 million people attending events.
For what it’s worth, the company insists it doesn’t keep the data it collects, and it doesn’t even take identifiable images of people. Sounds reasonable.
5. An Airport Vending Machine Used Facial Recognition To Dispense Coffee to People Who Yawned
Most people seem leery of facial recognition at first but there have been efforts to convince people to embrace it and they’ve worked well. All you really need to do is give people a reason to, and that’s what South African coffee company Douwe Egberts did.
The company set up a coffee vending machine in a busy airport but it had no way to actually pay for the coffee. However, cameras in the machine identified people who yawned and dispensed free coffee for them. Soon enough people figured out what triggered the machine and everyone who wanted free coffee could get it.
The whole thing was described as a marketing stunt, which it was, and stories detail the smiles and fun of all the people once they clued in. Less talked about was that the technology was conditioning people to produce a desired response, which worked out like gangbusters for the company and proved to them they can use facial recognition in campaigns to ensure customers are doing what they want.
4. There Are Hairstyles That Trick Facial Recognition
While makeup has proven capable of confusing facial recognition it’s not the only aesthetic choice you can make to throw the software for a loop. There are certain hairstyles that have proven confusing in the past as well, notably CV Dazzle.
The technique was designed by a man named Adam Harvey as part of his Master’s thesis and showed that combinations of hairstyles, makeup and even clothes could camouflage a wearer from what, at the time, was the most widely used facial recognition software algorithm.
While the designs might look eccentric to a living observer, it’s unlikely you’d ever think the person was actively trying to dupe facial recognition software. Instead, they’d just look like they were on their way home from Fashion Week.
While the designs might look eccentric to a living observer, it’s unlikely you’d ever think the person was actively trying dupe facial recognition software. Instead, they’d just look like they were on their way home from Fashion Week.
In much the same way Juggalo makeup could throw off the software, this was a less ostentatious way of doing it, or at least a more fashion forward way, that could include streaks of makeup or even hair obscuring part of a face, in particular one eye or in a way that alters the perceived elliptical shape of the face and head..
3. Facial Recognition Is Used in China To Stop Toilet Paper Thieves
When it comes to cutting edge uses for facial recognition, no one is making more of it than Beijing’s Temple of Heaven which is a sacred site and tourist attraction. In 2017, visitors using the bathroom at the Temple found themselves with facial recognition toilet paper dispensers.
The machine scans your face before dispensing a length of paper. The reason? People kept visiting the bathrooms to steal toilet paper. Now it gives you what it thinks you need, a strip of paper two feet long, and if you come back for more, it will recognize your face and refuse to give up the goods.
The machine scans your face before dispensing a length of paper. The reason? People kept visiting the bathrooms to steal toilet paper. Now it gives you what it thinks you need, a strip of paper two feet long, and if you come back for more it will recognize your face and refuse to give up the goods.
So what happens if, in good faith, you need more paper? That’s too bad because you’re only getting two feet and you better make it work. There is an inconvenient work around, if you have some time on your hands. The machines are on nine minute timers, so if you want a second length, wait the 9 minutes and you get more. It’s inconvenient for anyone who needs it but probably even more so for would-be thieves.
2. Malls Use Facial Recognition to Gather Biometric Data on Shoppers
We mentioned malls earlier when talking about Taylor Swift so let’s dig into that a little more deeply because this is far more pervasive than you may realize. Have you seen one of those mall directory kiosks that lets you look up maps of the mall and specific stores? There’s a good chance if your mall has one, it’s also running facial recognition software in the background with a camera trained on you and everyone else passing by.
A couple of Canadian malls got in trouble for this in 2018 because customers had not been informed that they were being photographed or for what reason. Two years later an investigation determined that the company that owns the mall had used the same technology in a dozen malls across the country to collect the images of 5 million shoppers.
The company claims it was only tracking foot traffic and nothing identifiable about the customers – just age and gender information for analysis. But it also collected video and even audio footage which officials claimed was just part of a test phase.
The mall says they informed customers because stickers on doors say cameras are being used for “safety and security” but the privacy commissions investigating didn’t think that counted when the actual cameras were harvesting biometric data to guess ages, genders and shopping habits.
1. Chinese Police Use Facial Recognition Sunglasses
Straight out of sci-fi, or failed ideas like Google Glass, police in China were outfitted with sunglasses in 2018 that had facial recognition tech integrated into the lenses. A camera in the frames can scan through crowds of people and using the same basic tech you might use to unlock your phone, the faces can be matched to a database. If there’s a face on file belonging to a wanted criminal, someone using a fake identity or anything like that, the glasses can alert the officer right away.
According to police sources in China, the technology proved to be very successful and aided in the capture of more than half a dozen individuals just by having cops wander through busy train stations.
It was also suggested that perhaps the technology could identify not just criminals but political dissidents or even just to profile people but surely no police force or government would ever do that.