Technology, cybersecurity solutions and related security software are in a constant mode of adaptation. To some extent, the adaptation is about staying ahead of the competition as well as the cybercriminals of the world. In other situations, however, the adaptation is because of an immediate need. Facial recognition software faced, and is continuing to face, a need to adapt as a result of a piece of cloth. More specifically, face masks/coverings!
The Process of Data and AI
Facial recognition is the process of using data and artificial intelligence to identify a person based on specific facial characteristics. The distance between the eyes and the position of the nose, chin, and mouth in relation to each other. All the data points are in facial recognition software. Once the specific measurements are captured, they are compared to other photos or videos that have already been gathered. But, as the use of face masks has increased and continues to be recommended, facial recognition software has far fewer data points to collect.
If you think about it, they lost the vast majority of their potential data points. And, as these data points (a.k.a. facial features) are covered up, the technology suffers. In some cases, facial recognition software failed to identify individuals with masks on, up to 50% of the time. In addition, some facial recognition algorithms struggled to even find a face when a face covering was hiding too much of the face.
Facial Recognition Software
Enter the adaptation stage of facial recognition software. Organizations have realized this and are testing adjustments to their software. Many facial recognition software organizations are working on technology that leverages the eyes, eyebrows, hairline, and general shape of the face. This will help to identify an individual. That may seem logical, but that requires not only an algorithm shift, but it may also require a new database of photos and videos of people with masks on for the comparison process. And, this may take time to build up.
So, what’s the big rush? The rush is based on security needs. Some airport security checkpoints, police departments, and drug stores began using facial recognition technology for security reasons. And, with masks blocking the ability for the technology to do its job, those organizations are at a loss. In addition, more businesses may look towards contact-free security options in the future to limit close contact security situations. All of these factors are creating a need for change, and one that doesn’t seem to be going away anytime soon.
Change, however, doesn’t come without a few bumps along the way. Facial recognition software was already under scrutiny by privacy advocates, and there was some indication that the software favored certain skin tones over others, causing concerns about racial biases. Will less facial data points cause even more identification issues?
Although we’re in a bit of a wait and see situation with how facial recognition will adapt to face masks/coverings, we can be sure of one thing. Without adaptation, technology cannot survive. The fact that numerous facial recognition organizations have already begun to explore new algorithms leads us to believe that some form of facial recognition software is here to stay. And that adaptation is a constant.