Let’s talk about face detection

Face detection, also known as facial recognition, is often taken for granted by us, just as many other technological achievements that we’re now able to access on a day to day basis, only a tap or a press away. Things quickly cease to be astonishing, and we hardly ever take a moment of our time to stop and acknowledge how far we’ve come, and how incredible is the technology that we carry in our pockets, for example. All it takes is a few seconds of lifting up our smartphone and facing the camera, and our faces will automatically unlock our device, just like that. Not so astounding, right? Well, it sure was in the 1960’s for Woody Bledsloe and his team, consisting of Charles Bisson and Helen Chan Wolf. Their work wasn’t recognized at the time because they worked for an anonymous intelligence agency (Source: Recfaces), but their work was incredibly innovative. They developed a system that could classify photos of faces by hand using what’s known as a RAND tablet, a device that people could use to input horizontal and vertical coordinates on a grid using a stylus that emitted electromagnetic pulses. The system could be used to manually record the coordinate locations of various facial features including the eyes, nose, hairline and mouth (Source: FACEFIRST). The accuracy of this technology was further improved in the 70’s, 80’s and 90’s, when features such as lip thickness and hair color started being taken into account (More features=more specific!) and even linear algebra stepped into the game. These were the beginnings of face detection technologies and it’s really important that we understand them before time-travelling to our modern days. Face detection saw exponential growth from the 2000’s until nowadays, and the proof is in the reach it currently has in our daily, collective lives. But how does it work? Fortunately, you don’t have to be an engineer or a rocket scientist in order to understand it. First, your face is detected (duh!) and analyzed by the device. During this analysis, the face is sorted into distinguishable landmarks, also known as “nodal points”. They’re like universal coordinates of what makes a human face distinguishable, such as the distance between eyebrows, the depth of eye sockets, the length of the nose, the thickness of the lips. After this, each nodal point is converted into a number in a database, creating a numerical code better known as a “faceprint”. Like a blueprint of your face, doesn’t it sound cool? Finally, your faceprint is compared to an entire database of different faceprints and a match is identified for your exact facial features. These technologies have become so intricate, that your phone could easily distinguish the faces of two identical twins and it could deny access to someone with a mask or makeup in an attempt to mimic another person. It’s almost 99% foolproof. Face detection is being implemented more and more with each passing day, and you don’t need to look too far to find it. Just open your Instagram app, slide right to open the camera within the app and choose a face filter! What you’d be witnessing is sophisticated facial recognition, where the nodal points of your face are instantly identified and tracked, so that a third party filter can adjust itself according to these points. It’s a really interesting integration between AR technology and face detection technology. Another cool integration between AR and face detection can be seen in video conferencing, where you can modify your background in real time or touch up your face, thanks to your nodal points. Speaking of touching up your face in real time, a fascinating app that makes great use of AR and face detection is FaceApp, which you’ve probably either a.) Played around with, or b.) Been annoyed by after seeing all your friends post it on their social media stories. What it does it quickly create a faceprint using your phone’s front camera and accurately modifies your features, making you look either older, younger or even of the opposite sex.

Facial recognition is even used for security purposes on a larger scale, as seen in the U.S, where the police have implemented these technologies in order to identify and track criminals, or in commercial surveillance, where shoplifters could be caught red-handed. Like all technological advancements, face detection has led to the rise of questions concerning its potential violations of privacy. Although these concerns are very valid, we shouldn’t lose sight of all the benefits face detection has reaped for us and continues to do so in our daily lives.