Facial authentication - Authenticates every skin tone and shade reliably

Identity and Authentication with Intel RealSense ID

For most of us, on any given day, we engage with multiple different authentication systems. To leave your home, you likely lock a door behind you with some kind of key. You unlocked a vehicle with perhaps a different kind of electronic key. You swiped a badge or clocked in with an employee card, and when you started work on your computer, you used a number of different alphanumeric passwords  to get access to the software, accounts or websites you need. These authentication methods rely either on a physical item like a key, or knowledge of passwords and pin numbers. In some cases, it might even be both, with two-factor authentication (or 2FA) requiring you to validate a login with a text message or app on your mobile device. 

All of these methods individually are reasonable, simple to use and relatively secure. The major failing of both of these types of authentication however, is the level of complexity which increases with every separate system you try to authenticate with. For every additional door that requires a physical key, the weight of objects you carry increases, or the chance that you will lose one either accidentally or because someone stole the item becomes a greater risk. For memory based systems, inevitably people end up with a small number of passwords that they reuse over and over for different accounts due to the impossible complexity of remembering a unique password for every single system that requires one. 

Why Biometrics are great for authentication

Biometric authentication refers to any type of authentication which uses unique physical or behavioral characteristics – such as a fingerprint or voiceprint. The advantages of biometric systems are many – unlike a physical key, you will always have it with you. Unlike a short password, a biometric key has magnitudes more information contained within your personal information, and can’t be forgotten easily. Stealing and reproducing your biometric information is also more challenging than with more traditional methods of authentication.

How it works

Any biometric authentication system requires some kind of sensor which takes a particular type of biometric data and converts it into a digital signature. The data is matched against the existing biometric data in your file to verify your identity. The most common example of biometric authentication would be using your fingerprint to unlock a smartphone.

Some types of biometric identification

Popular types of biometrics used for authentication

  • Fingerprints – as previously mentioned, most people will be familiar with fingerprint identification. While fingerprints are easy to capture and verify, there are some issues with using them as a general authentication system. Firstly, we leave fingerprints everywhere we go, which means that it’s fairly easy for a bad actor to acquire our fingerprints. In general, it’s best used in a situation that requires a second factor for authentication, such as the ownership of a phone or other device. With increased concerns about physically touching devices and sensors for hygiene reasons, as a broad authentication system fingerprints aren’t really suitable.
  • Voice recognition – each individual has physical characteristics which determine the sound of our voices, along with behavior characteristics which determine the way we speak – accent, cadence, tone and more. By combining the data from these two separate characteristics, a unique voiceprint signature can be created. While some systems are working on ways around this, research has shown that recorded voice samples from something like a YouTube video could be used to falsely authenticate someone.
  • Retina scans and iris recognition – both of these methods capture unique structures within the eye and use those to generate the biometric signature key. By adding requirements such as blinking to these systems, they are harder to spoof with photos. While both of these methods are relatively hard to crack, for the user they can be harder to operate. Close capture of the eyes means more time for authentication when compared with other methods.
  • Keystroke dynamics and signature recognition – both of these methods rely less on a specific physical characteristic, and more on behavioral ones to generate the authentication keys. How fast or slow you type or write, what rhythm, pressure or direction you do so can be used to create a unique behavioral profile. These methods work best in a situation where you would already be typing or signing. Adding keystroke dynamics to a password based system can make it more secure. Signature recognition used to authenticate payments or bank transactions would be much more reliable than previously possible.
  • Facial Authentication – this method uses either traditional 2D sensors or more complex 3D sensors to compare an image of the user’s face to a stored image or set of reference values. While some systems rely solely on the 2D image of the face for authentication, matching against a stored 2D image can be fooled by a high enough resolution photograph. Additionally, some systems that may store the image of a user’s face directly can have privacy issues. The advantages of a secure facial authentication system are many – it doesn’t require touch or the ownership of a specific device, and it can work rapidly in a wide variety of use cases.

Intel RealSense ID Facial Authentication

Systems like Intel® RealSense™ ID offer a number of advantages over other biometric authentication types. Our faces are unique – the distance from different features to each other can be used to create a biometric key. Using the face for identification is a better user experience than most of the other methods above, requiring no physical contact and less time to unlock. 

More advanced systems use more complex sensors that create a 3D image or print of the face, which make it much more challenging to spoof with a 2D image or even video. Intel RealSense ID uses dual cameras with an infrared projector to allow for the creation of a complex digital key that represents the user’s face without ever storing their image directly. This digital key is then encrypted, and stored either on device or a remote user database, but on device encryption with a secure element means that reverse engineering a user’s faceprint key even from stolen or hacked data would be extremely difficult. 

The design of Intel RealSense ID’s system also make it very challenging to spoof even with a 3D printed mask. While it may be easy to get a photo or even multiple photos of a person, obtaining the right data to be able to match with what the devices detect is much more challenging. Additionally, Intel RealSense ID is designed to adapt over time, using each verified authentication to help build a more accurate digital key to match against, so changing hairstyles, weight loss or gain, different styles or beard growth doesn’t mean you are locked out of your home. 

Because Intel RealSense ID uses infrared sensors and emitters, it also works well in a wide variety of lighting situations, both indoors and outdoors even in bright sunlight or near darkness. With a wide vertical field of view and the ability to recognize you from many different angles and positions, it’s a system that works well for everyone, everywhere. 

Intel RealSense ID has some similarity with fingerprint authentication in that it requires a unique physical device to be used though each device can support hundreds or thousands of users. This makes it suitable for applications such as banking or gate access control. Your bank could allow you to enroll for ATM access in a matter of seconds and that enrollment could then allow you to securely use any ATM they owned around the world without the need for a debit card or a pin number. Additionally, because the system adapts to changes in your appearance, no matter what your look today, you will still be authenticated accurately tomorrow or years from now.


Learn more

to get blog and news updates.

You may also be interested in

Let’s talk about how Intel RealSense computer vision products can enhance your solution.

By submitting this form, you are confirming you are an adult 18 years or older and you agree to share your personal information with Intel to use for this business request. You also agree to subscribe to stay connected to the latest Intel technologies and industry trends by email and telephone. You may unsubscribe at any time. Intel’s web sites and communications are subject to our Privacy Notice and Terms of Use.