Facing facial recognition

Since biometry has been becoming more mature, biometric authentication meant fingerprinting for most users. Then, face recognition became more reliable and more widely spread, for instance, as a means of unlocking one’s handset and, now, the face is increasingly used as a means of authentication.

One-to-one authentication, or in other terms controlling that the candidate face matches the one recorded in a database at enrolment stage does not create any issue to anyone. Users just discuss the convenience of using their face rather than their fingers in the context of authenticating themselves to their smartphone.

But the massive progress of computing power leads to use face recognition as a primary means of identification. In other terms, the face becomes the main index in a people’s database. Many applications are coming up, at the same time as many concerns as demonstrated by civil rights organizations.

Regulation needed in the US

Face recognition is such a powerful technology that it leads blue chip companies to ask for a code of ethics before any large size deployment. Microsoft and Amazon Web Services (AWS) are asking US lawmakers to set up a framework covering how face recognition should or should not be used. For instance, "when facial recognition technology is used in law enforcement, human review is a necessary component to ensure that the use of a prediction to make a decision does not violate civil rights," says Michael Punke, Vice President, Global Public Policy, AWS. The core issue revolves around face recognition in public or commercial settings and in connection with massive video monitoring.

The demand for regulation also comes from legislators as, in Massachusetts, a bipartisan group of lawmakers is introducing a proposal under which it would be illegal for any state entity to "acquire, possess, access, or use any biometric surveillance system." Also, in San Francisco, Aaron Peskin, a member of the city’s Board of Supervisors, proposed to ban the city agencies from using facial recognition technology.

UK police trials face recognition

The UK has made the choice to install surveillance cameras in public places for years leading to the installation of 4 to 5.9 million CCTV surveillance cameras in the UK, according to data from the British Security Industry Association (BSIA). Now, Automated Face Recognition (AFR) and artificial intelligence are increasingly used to process the huge quantity of images harvested from these cameras. This leads to questioning whether these technologies are becoming “demonstrably and routinely more reliable than a trained human operator undertaking the same task even in complex scenarios,” according to the recently published annual report from the UK Surveillance Camera Commissioner.

Already, the London Metropolitan Police has reported having arrested 3 people in a single day trial of live facial recognition in Romford in East London on January 31st. However, 2 out of the 3 arrested people have been released with no further action. The Metropolitan Police used NEC’s NeoFace technology, which scans the faces of people who walk by and alerts police to specific matches on a watch list. The police also stopped people who covered their face in the test area, leading civil rights group Big Brother Watch to complain and to deem the police action as “oppressive” and the use of Automated Facial Recognition by the police as “unregulated and legally questionable.”

At the same time, UK academics Joe Purshouse and Liz Campbell in an article in Metro, are calling for more legislation considering that “facial recognition technology is controversial, with research showing that it can be inaccurate and discriminatory.” They add that, when facial recognition has been used in large events such as the UEFA Champions League final in Cardiff, false positive matches outnumbered successful identifications. While the police claims the use of facial recognition is “lawful and proportionate,” this assertion has not yet been validated by the Courts.

China: face to pay

Meanwhile, in China, a shopping street in Wenzhou, Zhejiang Province, has become the first in the country to extensively apply facial recognition technology in making payments. Around 20 stores in Wuma (Five-horse) street has been designed as the test bench for Alipay “Smile to Pay” or “Dragonfly” technology that uses face recognition to authenticate users and complete a payment. Alibaba, the mother company of Alipay, say they are ready to roll out the technology in environments as diverse as hospitals, shops, restaurants and grocery stores. According to Xinhua Finance Agency, supermarkets, which have been equipped with the new payment equipment, have proven that they could save customers up to 50% of their time in queuing up. As of now, the cost of one Dragonfly payment equipment system is comparable to having three cashiers. Alibaba anticipates that facial recognition payment system is bound to be more convenient and safer than the password-based payment system and with lower costs.

More at the MWC

In the West, the next showcase for facial recognition technology is to be the Mobile World Congress where Comba’s ScanViS ID facial recognition (FR) technology will be used to expedite venue access at MWC19 in Barcelona, Spain, where around 107,000 attendees are expected. The solution will facilitate the registration process by allowing attendees to register and upload their photo on the organizer’s website before the event. After onsite passport/ID checking, attendees will be identified by the FR system and can move through the event access lanes without having to show their badge or ID to enjoy a seamless entry experience by just looking and passing. According to Comba Telecom, “ScanViS ID facial recognition solution can complete verification within one second on average, which provides a fast and security-focused access control management to effectively process large number of attendees.”