Face recognition is increasingly being used for biometric identification, from unlocking a phone screen or a banking application to access in buildings or check-in flights. Artificial Intelligence integrated facial recognition is used to track citizens, criminals, establish digital identity or biometric authentication.
A major thrust to this technology has come from governments across the world, esp. China and the US for extensive public surveillance and security. They are spending a large amount of money for integrated applications with existing citizen databases and to improve accuracy of face recognition algorithms and reduce false matches.
The use of facial recognition technology is used across verticals, such as BFSI, retail, healthcare, private companies or educational institutions, safety and surveillance of places including airports, railway stations, crossing state or country borders, etc. In the war against COVID-19, several governments have implemented these new surveillance devices in healthcare, public safety, and surveillance applications. This kind of “contactless” identification verification technologies have become of prior importance, to avoid transfer of virus or bacteria with biometric fingerprint scanners or other touch-based identity verifiers. During COVID-19, facial recognition had some major surveillance and authentication use cases.
Facial recognition systems can be used to accomplish one of two different tasks:
Verification (One: One matching) is used to confirm that the person is “who he claims to be”
Identification (One: Many matching) is when an unknown face is compared to a large database of known faces to determine the unknown person’s identity.
Face recognition uses AI technology to automatically detect the face from thousands of photos and videos, it can identify a person from the real-time footage across locations. It analyses and identifies the face through a photograph or video.
Every person’s face is broken up into numerous data points, it can be the height of cheekbones, the distance between the eyes and mouth. It searches data points and tries to account for variations. If a person is wearing a cap or sunglasses it can even be recognized by AI facial recognition. Facial recognition systems vary, but in general tend to operate as follows:-
Detection of Face
It recognizes the face and locates the image of the face then save it in the database
Analysis of Face
The software of AI facial recognition technology reads the geometry which includes distance between eyes, the depth of eye orbit, distance from forehead to chin, the shape of cheekbones and outline of lips, ears and chin. It is used to distinguish a face from others to identify the facial landmarks.
Changing the image to data
The captured face transforms the face into digital information based on the person’s facial features. The analysis is turned into a mathematical formula. This code is becoming a faceprint which is as unique as a thumbprint.
Finding a match
To find a match, the faceprint is compared with the database to recognise the faces.
Smart Phone Safety
In today’s time, every smartphone has a facial recognition verification feature that helps to unlock the devices. It can be used to scan the face to unlock different applications on the phone, to eliminate the need to remember passwords. Hence, it offers safest means to protect unwanted access to applications, even their data and guarantees its safety, even if the phone is stolen. Weak passwords are much easier to crack or guess if the phone is stolen.
Along with extensive CCTv monitoring, face recognition technology helps in registration, recognizing the patients, and even to detect emotions or pain. Apart from regular surveillance and asset monitoring, it can be increasingly utilised for Patient monitoring, health-staff identity, support pain management procedures and track patient medication consumption.
Busy Airports have millions of visitors moving every day. “Smart gates” can continuously monitor visitor faces and movements of passing visitors and can reduce time for scanning with combination of facial recognition and other techniques. Integrated with central databases, these checks can “trigger alerts” to airport authorities, for any suspected individuals or criminals on such gates. Thus, it can boost security, reduce the waiting time for immigration checks and bring efficiency.
Banks can increase safety of its cash in ATMs, branches, locker rooms, identification of its customers, as well as their employees with the help of CCTv surveillance and AI based facial recognition. Mobile banking applications are using face recognition to conduct remote KYC check of their customers. They can also use it to restrict unauthorised access of banking areas.
In retail, face recognition can be used identify staff, unauthorised access or alert shoplifting. It can be used to capture images/video for a customer’s demographics or interest profile, their shopping journey. Thus, it can give valuable insights for marketing to drive promotions/offers and identify repeat customers across different stores to improve their customer experience.
On government level, facial recognition can help to identify wanted criminals and terrorists or finding missing children. Facial recognition technologies have been deployed via homeland security, federal and regional law enforcement departments, and other security agencies across the globe.
AI facial recognition can ‘do more’ to provide a better convenience. For example, it can be used to buy goods and “Fast Check out” from retail stores without paying with cash or credit card. A faster access for office premises by employees to account attendance or authorise candidates at online testing centres, lifts/hotel stay access or while entering through airports/ railway stations, hospitals etc.
Contact less Authentication
AI facial recognition can augment security guards at entrances, to perform fast authentication of visitors with accuracy. The technology collects a set of unique data pertaining to an individual associated with their facial features and expression to identify, verify and/or authenticate a person. During pandemic, FR proved to be more dependable, as it can analyse the person’s face from a safer distance and verify identity of people through contactless gates.
Threat for privacy
This is one of the main concerns that people feel is a breach of their privacy, the majority of facial images are taken without people’s consent.
Facial recognition technology is omnipresent in every video surveillance camera, which can help the government to track wanted criminals. But it even tracks down ordinary and guiltless people at all time without their consent.
Possibility for error
Every technology has a possibility for an error, even facial recognition is not error-free. It can lead to people who are wrongly accused of crimes they have not committed. In the year 2018, Amazon face recognition matched with 28 Congress members as criminals.
Face recognition is an extremely powerful technology if it is used properly. For example, Snapchat uses facial recognition technology to uses filters that are mould into user’s faces, Google photos uses to sort pictures and tag the person recognized in photos. Real-time face recognition is most accurate when faces are compared in a static image, but its accuracy drops when photos are taken in the public. AI facial recognition can transform the whole video surveillance system. However, most leading companies have expressed their concerns over misuse of this technology and a need to regulate.
Dhaval brings over 22 years of experience in Product Management and Customer Engagement. His deep understanding of customer pain & technology led to the launch of multiple products that scaled to meet the needs of some of the largest global telcos & resulted in 10x growth for Elitecore. He was a founding team member of Elitecore (now STL Software), he built teams and led 700 employees to adopt successful agile product delivery practices.
Passionate about AI & Computer vision, Dhaval founded AIVID with the vision to solve critical but repetitive tasks using automation leaving humans to focus on value creation.