Due to the Covid-19 outbreak, people all over the world, now wear masks on a daily basis whether they are on road, working, driving, and sometimes even at home. But how has this ‘new normal’ affected the facial recognition systems that we use for identification?
Hassan Ugail, Professor at the University of Bradford, seemed to notice the glitch when he tried to unlock his iPhone via facial recognition.
I kind of have to take my mask off. I would rather it let me in by just looking at my eyes.
This problem is not just true for only iPhones but also for many commercial systems that authentic people using their facial features. The issue became public when the National Institute of Standards and Technology (NIST) in the US. issued a report in July 2020 stating that ‘ error rates of 89 different facial recognition systems have increased from less than 1% to a whopping 50% when the mouth and nose or bottom half of an individual’s face was obscured’
So what is the solution?
In his research last year, Hassan along with one of his PhD students had found out that only half of the face is enough for facial recognition algorithm to work. Many firms are now working fast to update their systems so that they can recognize people with masks where facial recognition is used to authenticate access to an individual in high-security areas.
Mei Ngan, a computer scientist at NIST says,
Developers are indeed adapting their algorithms to handle face masks
Progress is being made and the facial recognition software are being trained to cope with different size and shape masks as well. To do this, a variety of superimposed shapes to represent multiple sizes of masks are used rather than pictures of people wearing masks.
As ‘contact free’ technology becoming popular during the pandemic, companies are using this point for their marketing needs, telling people how facial recognition is ‘a contactless, potentially more hygienic, means of verifying identities in public places’.
A Swiss Company has declared about their Tec5 Algorithm that its working just fine even if people were wearing masks. That is why Tec5 based systems are being used in many Schools and factories of South East Asia to mark daily attendance.
Co-founder Rahul Parthe says that the algorithm is reliable,
If somebody was wearing a mask and sunglasses and had a hat then you were basically missing a lot of their faces. That’s where we had to put in some requirements where a person would have to remove their hat or mask
Tech5’s systems can also do a headcount to see how many children are present in one day. Moreover, they can passively monitor images from a CCTV camera in a corridor and show results. However, pupils are advised to not wear hats, otherwise, the system might get confused.
Another company called Trueface have accepted how covid has forced them to do changes in their algorithms:
The onset of Covid had really forced companies like Trueface to rethink what areas of the face are more important for facial recognition.
However, Trueface staff were quick to do the changes and had started system updates as early as February 2020, when the spread of the pandemic became serious outside China. Company Founder Mr. Moore believes that the covid 19 will ultimately increase the rollout of facial recognition, particularly where authentication is needed.
In October 2020, his company announced it had secured a contract to install facial recognition access control systems (using touchpads and surfaces) at 2 US Air Force bases to reduce contact between people at the bases. Mr Moore thinks that this is just the beginning and very soon venues more public places such as sports stadiums will also turn to ‘contact-free’ identify verification. He says,
Facial recognition becomes the way to do that without having to hand over a key card, badge or ticket, it’s reducing the contact you would have as the visitor, employee or spectator. As that worker, you don’t want to be touching a thousand cards
It’s important to note that, while basic hygiene and washing of hands remains crucial in the pandemic, the US Centers for Disease Control and Prevention has declared that although the virus spreads easily from one person to another, it won’t really transmit via contact with a contaminated surface. Nonetheless, the outbreak of this deadly pandemic is clearly not only creating new challenges for the providers of facial recognition systems but lucrative opportunities as well.