Tapping into the potential of multi-modal biometrics

24 February 2021

Tapping into the potential of multi-modal biometrics

24 February 2021

Tapping into the potential of multi-modal biometrics

Biometrics are fast, easy, and eliminate our reliance on passwords and PINs, however, despite their growing popularity and benefits, biometrics have their own unique challenges when it comes to widespread adoption.

Most solutions on the market today only offer single-modal biometrics to authenticate users. This could be one of the reasons that the adoption of biometrics has been slow, particularly at the enterprise level in industries that would benefit most from the technology in terms of user experience, but require greater assurance that their users or employees are who they claim to be.

Advancements in biometrics have reached a point where the technology can reliably replace passwords, eliminate threats, and enhance the customer experience all at once. It’s time for enterprises to take a multi-modal approach to biometric authentication and identity management.

What is biometric security?

Biometric security is a method of authentication that uses unique physical or behavioural characteristics to verify a person's identity.

Biometric security is already in use, on a consumer and enterprise level. Common examples include unlocking devices using a fingerprint or face scanning, and they’re increasingly used in applications such as banking, border control, and access control systems.

There are two main types of biometric security: physiological and behavioural. Both physiological and behavioural biometrics have their advantages and limitations, and may be more or less suitable depending on the specific application and security requirements.

An unimodal biometric security system will only use a form of biometric to allow the user to gain access. Whereas, a multi-modal biometric security system will use two or more biometric identifiers.

Some emerging biometric security methods include a wider range of biological traits, such as iris or retina patterns, voice, or even a person's typing rhythm.

What are physical biometrics?

The most popular biometric solutions leverage a user’s unique physical traits to authenticate them. Biometric templates are created by mapping a user’s unique physical traits. For example, at Keyless we use Neural Networks, a deep learning technique to extract high-dimensional feature vectors from sequences of images of the user’s face.

What are behavioural biometrics?

Behavioural biometrics, on the other hand, leverage deep learning to recognize unique patterns in how a user interacts with a device over time. These patterns could be based on how the user taps, swipes, or holds their device. Once patterns have been recognized, they can be transformed into biometric templates that can be used to monitor access to a user’s device in real-time.

Behavioural biometrics offer enhanced usability as they require no conscious effort from the user, allowing for continuous, frictionless authentication, enabling companies to frequently verify that a user is who they say they are, without disrupting the user. Physical biometrics, on the other hand, are better suited to one-time authentication, as they typically require conscious effort from the user.

Why multi-modal biometrics?

The reliability of unimodal biometric systems plays a part in this. Unimodal (or, single-modal) solutions are

Less likely to withstand advanced spoofing threats.
Less likely to correctly match biometric templates as a user’s physical characteristics change, whether due to the natural ageing process or suddenly as a result of an accident or disability.
More likely to be biased, or less effective at recognising demographics other than white men, particularly for demographic subgroups, such as black women.

As a result, the reliability of single-modal biometrics is often questioned.

Multi-modal solutions that combine physical and behavioural biometrics enable companies to enhance security, without disturbing the user experience.

Multi-modal systems have the power to

Lower the false rejection rate, because they have a higher accuracy than looking at a single biometric trait alone.
Prohibit hackers, as faking two unique biometric traits is much harder than faking one.
Combat noise in the data, by offering users a more flexible approach to biometric security.

Frictionless security powered by the user’s identity, not what they know

When it comes to enabling fast, secure authentication in the workplace, biometrics are inherently user-friendly, and a multi-modal approach to authentication can directly solve the trade-off between convenience and security.

Multi-modal solutions give companies greater control over managing remote access to their private systems and data. For example, if an employee logs in, but then the system recognises unusual behaviour (such as different keystrokes or tapping patterns), then it can log the user out.

The Keyless approach to multi-modal biometric authentication

At Keyless, we take a multi-modal approach to biometrics which allows us to leverage both physical and behavioural biometrics to offer a seamless multi-modal authentication solution. With this approach, it allows companies to reliably identify users.

The Keyless protocol allows for several physical biometric modalities including facial, fingerprint, voice, and iris recognition; as well as behavioural modalities, like keystroke and swipe recognition.

How it works: using deep learning to ensure high-accuracy

To offer behavioural biometrics, we must first learn how users interact with their devices. At Keyless, we do this by feeding deep-learning algorithms unique, user-generated data captured through sensors on the user’s mobile device.

Each input can be gauged on its reliability in authenticating an individual user. Once it’s considered reliable enough for identifying the user with high accuracy, the feature can be added to the user’s multi-modal biometric templates.

Ensuring privacy compliance

Since biometrics are uniquely linked to the individual, it’s important that companies processing biometric data take extra measures to protect it against both new and emerging threats

“Biometrics have the potential to make authentication dramatically faster, easier, and more secure than traditional passwords, but companies need to be careful about the biometric data they collect.” — Maria Korolov

Keyless ensures that the biometric data we capture is never at risk of being stolen, compromised, or lost by combining multi-modal biometrics with privacy-enhancing technologies. The combination of privacy-enhancing technologies, deep-learning, and multi-modal biometrics uniquely addresses the greatest barriers facing the adoption of biometric authentication solutions.

To gain a better understanding of how privacy-enhancing technologies can be leveraged to protect biometric data, you should read the following articles:

Closing the gap between security and convenience with multi-modal authentication

By presenting multiple biometric challenges to users, multi-modal solutions protect private systems and data in the rare event of successful spoofing attempts or account takeovers that take place after the point of authentication.

Thus, multi-modal biometric solutions can help companies close the gap between security and convenience, allowing for more seamless authentication and identity management experiences that enhance security and privacy compliance, without sacrificing the authentication experience or productivity.

Interested in trialling Keyless to enable multi-modal biometric authentication?

Keyless™ multi modal authentication can help deliver secure and seamless digital experiences for your end-users and for remote workforces.

If you're interested in learning more about multi modal biometrics, or if you’d simply like to learn more about our technology, then please feel free to get in touch with our team.

You can email us at info@keyless.io

We’re always keen to have a chat about how we can help businesses on their journeys towards multi-modal biometric authentication.

Tapping into the potential of multi-modal biometrics

24 February 2021

Tapping into the potential of multi-modal biometrics

24 February 2021