Joe Pinkstone | The Daily Mail | Source URL
Chinese companies specialising in facial recognition have upgraded the technology to identify people wearing masks in the wake of the coronavirus outbreak.
Chinese company SenseTime, described as 'the most valuable AI startup in the world' and worth at least $4.5billion, is one of several firms improving its facial recognition to ensure a person cna be identified without them taking off their mask.
Instead of relying on having to see a person's mouth, the system is able to learn a person's identify from just their eyes and upper nose region of their face.
People around the world are increasingly opting to wear medical masks and even respirators to prevent catching the potentially fatal COVID-19.
The disease has claimed the lives of more than 2,700 people around the world and infected a total of around 80,000.
Researchers have been working on expanding facial recognition's capabilities for several years, making it possible to identify people in glasses, wearing makeup, and with facial hair.
But the outbreak of coronavirus in Wuhan, which has now spread around the world and is thought by medical experts to be on the brink of a pandemic, has accelerated the technology's roll-out.
SenseTime announced earlier this month that it was incorporating new features to its already omniscient system to combat masks and coronavirus.
These include thermal imaging cameras to identify people with a fever, a key indicator of coronavirus.
SenseTime told Quartz this was currently live in three public locations.
It has now also added the mask-proof algorithm to tighten its vice-like grip on the personal privacy of civilians.
The system is widely used by Chinese authorities and can be used to keep tabs on people disobeying strict instructions that mandate residents must wear masks to halt the disease's spread.
'With these methods, management personnel are provided with a full range of epidemic prevention information such as body temperature, mask-wearing status, and employee identity,' the company said in a statement.
Stanford University postdoctoral fellow Amarjot Singh works on developing facial recognition using AI.
He told the South China Morning Post: 'Face recognition identifies a person by locating several key points on the face and connecting them together to form a unique person-specific signature.'
Traditionally, these include myriad markers around the eyes, nose and mouth.
But recent developments have meant a match can be made from just markers placed around the eyes and nose.
SenseTime's algorithm is designed to use 240 facial feature key points but can make a match using just the parts of the face that are visible outside a mask.
The company claims offices that use the system to identify and admit staff into the building will work without needing staff to take off their mask.
'The system that applies for office building entrances reads the facial characteristics that are not covered, including eyes, eyebrows and parts of the nasal bridge to identify employees,' the company told Quartz.
'So that the employees can keep their masks on at all times while still gaining access to their offices.'