Will Face Recognition Technology survive with Face Mask becoming the new Norm?

Facial recognition technique is an excellent method of enhancing security as well as enabling touchless access control within several facilities including buildings, in an era of social distancing. However, since experts are more and more recommending face masks to avoid the spread of coronavirus. Is it possible that the evolution of the face mask economy will affect the performance of facial recognition algorithms?

Recently, launched Google Pixel 4’s version that claimed of recognizing a person even when they were wearing a sunglass or shaved their beard. However, they never stated that if the technology will work be effective when wearing a face mask. Google isn’t the only technology giant that develops consumer electronics with face recognition technology but Apple is another one in the queue. In addition, Asians have always been complaining about face masks obstructing their face recognition, as wearing a mask is very common in Asia when people are sick. And now with COVID-19 outbreak making face masks, a new normal will this technology really be able to survive?

Probably no, because for face recognition algorithms’ to detect a covered face will require professionals to teach algorithms to read a face password with partial face information. This will mean that when a portion of the face is covered, then a facial recognition algorithm will have to focus over certain landmarks of the face it can view to determine a near accurate match. Moreover, it is a process that would necessitate training as well as several alterations for getting a right, yet flexible algorithm capable of adapting a partly occluded face but still able to deliver an exceptional result.

Nevertheless, this would actually be a complex process that would possibly require a number of variables contributing towards it including the quality of registered face image, lighting impact followed by several other environmental variables. Also, some people have especially unique facial traits, which might make their recognition via partial face information a bit difficult then several others whose either upper half traits would match or lower half.

On the other hand, this procedure might certainly hamper several security protocols and requires a real-time testing offering a technology that would be immune to masks.