What is facial recognition?
Facial recognition is one of the most controversial topics in the technology sector today. Since the technology emerged in the 1960s, this method of identity validation has grown consistently, in both its accuracy and its usage. In 2017, Apple introduced Face ID on iPhones, launching facial recognition to the mass market. By 2020, the global market value had reached USD $3.72bn. This is predicted to more than treble by 2026.
Who uses facial recognition?
There are two main types of facial recognition systems: one-to-one and public. The controversy surrounding this technology primarily applies to public facial recognition. These systems are for surveillance purposes. Cameras are set up in public places to scan many different faces at once and identify individuals within the crowd.
Public facial recognition
Casinos and airports might be the most likely places you would encounter a public facial recognition system. Additionally, some governments are now utilising this technology to track and police their citizens. China is the world’s largest user of such public cameras, to the extent that 54% of the world’s surveillance cameras are located there. More recently, Iranian authorities announced a plan to use facial recognition to enforce a law requiring women to wear a hijab in public places.
One-to-one facial recognition
The other type of facial recognition is a one-to-one system. This category covers most of the access control and building security solutions. Faces are scanned individually and compared against a pre-defined database of users. If the scanned face matches a saved face with the appropriate access rights, then entry will be granted. Face ID on an iPhone is an example of a one-to-one method.
The most common objections to facial recognition are concerns around privacy and consent. In most cases, these worries apply only to public systems. For example, privacy lobby groups such as Big Brother Watch claim that the use of facial recognition in retail stores violates customers’ right to privacy and breaches data protection laws. On the other hand, retailers such as Southern Co-operative believe the use of this technology will prevent violent attacks against staff and other customers.
In one-to-one facial recognition systems used for access control purposes, users have to be registered individually in the database. Upon enrolment into the system, each face must be scanned and saved against a user. While every organisation has different policies and practices for access control, one-by-one enrolment means the user always has an opportunity to provide or deny consent.
It’s always a good idea to have multiple methods of validating users for access control. A backup keypad code or swipe code comes in very handy in the rare case that your facial recognition camera fails. Or, if a user does not consent to their biometric facial data being kept in the system, a separate method can be provided. By combining the speed and convenience of facial recognition with clear consent procedures and alternative options for those who prefer them, access control systems can balance the best of both worlds.
How does facial recognition work?
Facial recognition technology uses the natural uniqueness of human facial features as a means of verifying personal identity. Specialist cameras scan faces to analyse and measure the shape and relative positions of key facial points. This information is assembled into a unique biometric data map of the face and saved into a database.
When a person next approaches the camera, their face is scanned, analysed, and compared to the database. If the face matches one of those in the database, then entry will be granted (provided that user has the appropriate access rights).
There are four steps to identifying faces:
1. Detection of the face, where the camera picks up a person in the vicinity.
2. Analysis of the face, where an image is captured and digitally examined.
3. Conversion of the face image to biometric data in the form of a facial map.
4. Comparison of the map with the existing database.
History of facial recognition technology
Early attempts at computer-driven facial recognition were made in the 1960s. This involved pinpointing the coordinates of certain facial features from a photograph onto a graphics tablet.
The coordinates were used to calculate different distances between the points, such as the distance between the inner and outer points of the eyelid. A computer would then compare the sets of distances to find potential matches.
From its beginnings in identifying people from photographs, the focus has now shifted resolutely to live facial recognition. Organisations are seeking to analyse and identify faces rapidly, in real time, and often from busy or low-resolution images. There are now a wide range of techniques and technologies used in facial recognition.
For example, Face ID works by projecting thousands of invisible infrared dots onto the user’s face. This generates a map of the unique contours of the face that can then be used to identify them.
The iface solution
iface is a biometric facial recognition solution that matches faces in just one second. Providing rapid, reliable access control, iface is faster, more efficient, and more accurate than the human eye. The device can be mounted on walls, desktops, posts, and turnstiles. In addition, iface comes with a built-in high security card reader as a backup or multi-factor authentication method. Choose any combination of facial recognition, RFID card or tag, and a PIN code to suit the needs of each application.
How does iface work?
iface is built with two separate cameras. One utilises standard white light to capture images of faces. The other puts out invisible near-infrared beams that map facial features. The combination of white light and near-infrared light generates a highly accurate biometric template of each face that is saved in the device’s database. As well as increasing the reliability and efficiency of iface™, the infrared camera means that faces can be scanned and validated even in low ambient lighting conditions.
Key features of iface
• Standalone facial recognition OR simple integration with access control
• Analyse and verify faces in just ONE second
• Built-in MIFARE DESFire card reader and keypad for multi-factor or backup authentication
• Memory for up to 20,000 face templates
• Anti-spoof fake face detection – a photograph cannot be used to falsely gain entry
• Tamper detection and international language support
Integrating iface with access control
iface integrates smoothly and efficiently with online access control systems. CDVI’s isync software is used as a mid-point, making it easier than ever to integrate. There are even pre-built integrations with Paxton Net2, ATRIUM, and other popular access control solutions. It takes just a couple of clicks to add facial recognition to any relevant user in the access control database.
For smaller-scale installations with no networked access control system, iface also operates as a simple standalone solution. Users and their face templates are saved on the unit itself, with quick and easy enrolment in just a few seconds.
Is facial recognition really more secure?
As with any security or access control solutions, there are pros and cons to facial recognition. One of the primary benefits of this technology is that it does not expect anything of its users.
With a keypad or a swipe card system, the user is expected to remember a code or carry a physical credential. All you have to do with facial recognition is turn up. You can’t leave your face at home. You also can’t lose your face or have it stolen; one of the biggest security breach risks with a physical credential system.
iface is equipped with spoof detection technology. Its near-infrared camera emits beams onto the face that map its contours, creating a 3D impression of the face.
That means that even if someone presents a photograph of a registered user to the camera, iface will not be fooled.
On the other hand, your biometric data cannot be changed or replaced in the way that a keypad code or swipe card can. Once that data is stolen, that’s it. The risks of a breach might therefore be lower, but the consequences in the case of one might be more severe.
Fortunately, not only does iface come with a tamper detection feature, face data is protected too. Even if an intruder was able to steal the unit from a premises, they would not be able to extract biometric data and use it to recreate actual faces.
The measurements taken upon enrolment of the face generate a facial map, which is in turn encoded through an algorithm. It’s not possible to backwards-engineer this data to recreate the original image.
There can never be a 100% guarantee of security with any access control system. Security installers and facilities managers must balance a range of needs: robust security, user convenience, administrative burdens.
However, facial recognition solutions such as iface provide a combination of benefits unlike any other solution. The inherent security of biometric data, the processing rate of one person every two seconds, and the simple, intuitive interface make for a powerful mix.
For more information, visit: www.cdvi.co.uk
This article was originally published in the October 2022 edition of Security Journal UK. To read your FREE digital edition, click here.