New biometric capabilities further extend flexibility of Aware’s Knomi authentication framework

The Knomi mobile authentication framework from Aware is well-established as a leading option for onboarding and biometric identity assurance in multiple verticals.

The flexibility to deploy Knomi in different environments and workflows, easily integrating biometric authentication with liveness detection and even multiple modalities is a key differentiator for Aware’s flagship technology.

Several new capabilities have also been added to Knomi over the course of 2020, including enhanced algorithms for passive liveness face and voice spoof detection, and voice biometric authentication, more features for on-device implementations with Knomi D, and upgraded accuracy for people wearing masks.

Face and voice can be used together for highly-secure multi-biometric fusion, but a greater number of organizations see value in having the option of one or the other in the same application, Aware Chief Commercial Officer Rob Mungovan tells Biometric Update.

Choice through framework model

The company currently has at least two customers using face and voice biometrics in a time and attendance workflow, he says, finding the extra modality makes it easier for their workers. They have kept the user’s PIN in the workflow, as is typical of Knomi implementations, he says, and makes it essentially a three-factor process.

“The first thing the system does is it goes to the index of that PIN or username, and says, ‘here’s the two samples that were just provided, do they match the enrolled samples and if so, what is the score,’” Mungovan explains.

As opposed to multi-factor authentication with non-inherent factors like physical tokens, which are either present or not, combining two biometrics allows false match rate (FMR) to be significantly lowered for maximum security.

The customers who have implemented voice for time and attendance found that it is easier for their workers to have the option.

“Voice is a somewhat weaker biometric,” he points out. “You can’t generate a signal that’s as strong as a good quality facial image, you don’t have as many features, but none the less there are certain use cases where voice works well, particularly when a good facial image cannot be acquired for one reason or another.”

There is also some talk among Aware customers of using voice biometrics to help pensioners, or “certain segments of the population that maybe are older or not as familiar with how to use smartphones, or how to present their face to the phone, the voice is being considered as a primary option to the face, with the face being secondary if the voice fails.”

Knomi’s text-dependent speaker recognition allows customers to set a pass phrase, and when enrolling typically ask for three samples of the same phrase between about five and ten seconds in length. In the company’s demos, it uses the phrase “Hello Knomi, please verify my identity.” However any phrase in any language can be used. After enrolment a single utterance of the pass phrase is sufficient for a strong authentication.

Initial customer interest remains primarily in face, Mungovan says, though he also notes that is “starting to get pockets of interest in voice. Certain regions, it seems, have expressed an interest in voice. Financial services express an interest in voice. But the adoption rate for unattended authentication lags that of face.”

Continuing to build out the flexibility that Aware provides to its customers is the key.

“The addition of another biometric that’s able to be collected from a mobile phone is a differentiator,” Mungovan explains. “A lot of our agreements, with a lot of the customers we have in place have a common theme of ‘I really like the face part, the voice part I think is of interest too. I’d like to be able to access that and utilize that under our agreement if in fact we encounter use cases where people are going to ask for it.’”

As companies increasingly plan for technologies that can enable different contingencies, there are very few biometrics providers that can offer both modalities.

Financial services continue to be the largest customer group for remote biometric authentication, but implementations for remote workforce management and unsupervised remote time and attendance tracking are also increasing.

Implement based on workflow, not other way around

As consumers and organizations alike become more used to and comfortable with biometric authentication through their mobile devices, potential customers coming to Aware tend to have a better understanding of their own requirements than the options available to them.

That applies to liveness detection, where passive liveness spoof detection is much easier to use and cannot be gamed by observing the mechanism, Mungovan argues. When Aware explains the differences, he says, customers are consistently won over.

On-device versus on-server is also important to consider, and a source of some misconceptions.

While there are several situations in which on-device biometric solutions like Knomi D should be considered, Mungovan says, as when connectivity is limited, more often there are advantages that customers can realize with server-based implementations.

“One is that your algorithms can get very complicated and you can put a family of algorithms on a centralized machine where you can’t really easily put them onto a mobile phone,” Mungovan says. “So you can just continue to build up your spoof detection capability by updating a server-based deployment. Secondarily, while we do provide very lightweight apps on the phone and sample apps and source to those apps so our customers can build their own branded app, and these apps do help ensure the best quality facial image is captured,” by performing auto-capture only when the image quality is optimized for the analysis algorithm.

There are other advantages, including for auditable records, and if a spoof does get through, you can figure out how and remedy it, Mungovan observes.

It’s also notable that Knomi provides a no-app option. A facial image can be captured with the native mobile browser, or a desktop browser, and submitted to the server for spoof analysis, enrollment, and authentication. “This is a super lightweight option,” states Mungovan, “Which some customer really like.”

Aware offers all three methods, in its quest to provide a toolkit that allows customers to add biometric authentication to their existing workflow, or plug it into an existing onboarding application. They can do so with RESTful services designed to be integrated into the IT asset of a large institution, rather than requiring a SaaS subscription or a “big monolithic piece of code.”

“One of the things that’s unique about Aware is that the design of our software implicitly is focused on enabling our customers to build out and maintain their own capability,” Mungovan emphasizes. “In other words, it’s a very flexible framework.”

By bringing the best of biometric authentication to customer workflows without imposing any compromises, Aware plans to secure the next wave of remote transactions with Knomi, while delivering the experiences that end-users and organizations alike expect.

Article Topics

Aware  |  biometric liveness detection  |  biometrics  |  facial recognition  |  identity verification  |  multi-factor authentication  |  onboarding  |  passive authentication  |  spoof detection  |  voice biometrics