The Effects of Glasses on Facial Recognition

Let’s talk about your face. Most of us have them, and as facial recognition technology advances, the concept of public anonymity will soon become obsolete. Depending on your personal beliefs, you may comfort or scare by this idea. But before considering two different approaches to this sector, we must first consider the impact of glass on this type of technology (if any).

First, there is a difference between face detection and face recognition, and in fact, glass is more likely to affect face detection.

Face recognition consists of determining from an image which the person really is. If you have a picture of a person, you want your software to be able to tell you the identity of that person with a high degree of certainty.

People who wear sports sunglasses pose challenges to both face detection and facial recognition software. Glass, especially reflective sunglasses, can prevent the algorithm from finding the reference point needed to determine if a photo has a face. If there is no face detection, there is clearly no face recognition.

Sunglasses trick advanced facial recognition software by changing a small bit of colored information on the face.

 Some people have deliberately tried not to be recognized by this type of software. Celebrities, in particular, tried to avoid being recognized. Some people are really worried about their privacy and don’t want their photos posted or seen online. These photos often contain metadata about time and place.

Benefits of Wearing Bright and Dark Sunglasses

Some people wear bright, dark sunglasses, which are often effective enough to stop recognition. Many face detection algorithms still detect them, but that’s not too much of a problem for celebrities who want to avoid recognition.

Each face detection and recognition product uses a different algorithm. Many of these algorithms recognize when people wear black sunglasses. Prescription eyeglasses with transparent lenses are generally not a problem for face recognition and detection software, as important eye details are still visible.

 If they cause any problems, the software will continue to inspect other parts of the face and remove the obstructed part from its analysis. It is not uncommon for faces to be detected and recognized with up to 30% occlusion.

Face recognition software relies on neural networks to compare the colors of pixels in specific areas of the face and match one image with another. When something (like these glasses) changes the color of this pixel, it confuses the software. Researchers say the 3D-printed glass is suitable for both phishing and evasion attacks. Spoofing “requires a face to be recognized as a particular another face”, but avoiding it is an attempt to avoid detection. Overall, the system achieved a 90% success rate in blocking facial recognition software.

But when people wear dark or bright sunglasses, especially those that hide the pupils of the eyes, the best algorithms in the world are still struggling. This problem is exacerbated when the glass obscures or obscures the distance between the eyes. This is an important part of many algorithms. Most facial recognition algorithms begin by first determining eye features and then comparing them to an existing image database before moving on to other facial features.

Leave a comment

Shopping cart