Computer

How does facial recognition know who we are?

One of the functions of AI is what we call computer vision, where the system does not generate graphics, but rather learns to identify elements in the environment and act accordingly. This was initially used to catalog cat photos as an experiment, and no, I’m not kidding. However, today many use their own face to unlock the mobile or make a bank transfer. It is easy and comfortable, that is why we are going to explain how it works.

Facial identification, how does it work?

What facial identification does is detect a series of elements of the face that are unique to each person and that will not be modified. To do this, it uses your PC’s webcam, although what is dangerous here is the fact that the system does not have the capacity to differentiate a photo from the real person. However, some of these systems use very poor quality biometrics by taking the entire face and not just the key parts.

What happens if the physiognomy of the face changes for whatever reason? The person has lost weight, wears makeup, has grown his beard, has changed the shape of his glasses, and so on. All this may seem trivial to you, but they are not and a bad face identification system can leave you without access to your data. So how can it work if we have continuous cosmetic and appearance changes in our lives?

facial recognition

Well, due to the fact that computer vision allows you to identify the basic elements of the face and with this, identify a series of points that will be used to determine the person’s face and compare it for facial recognition. It does this by not taking one image, but several of them in order to triangulate and have more precise information on the different points of the user’s face.

So it is important that the web or front camera of the mobile, responsible for the capture has sufficient resolution in order to have a good enough sample to make the corresponding comparison. The process is not easy and has not been made possible on mobile until neural processors, which are still unheard of on many PCs, have become popular.

And what about the use of the voice?

The human voice is so much easier to construct and fake, it actually takes very little time with today’s systems to create an entirely new synthetic voice or even mimic another person’s. Which is dangerous since one person can impersonate another and place them in a compromise or in a very uncomfortable situation.

However, the use of voice in combination with facial recognition can be used as confirmation and prevent someone from using a simple photo to access private content and not authorized by its owner. In any case, most of the built-in microphones stock in people’s laptops don’t have enough power to know how to differentiate nuances in audio.

Related Articles

Leave a Reply

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