Today, AimBrain offers five biometric modules, built and maintained in-house. Now, as there are a lot of different biometric options to authenticate with, I’m taking a different approach.
Here in part one, I’m painting the picture of the modules working as a team, under Jurgen Klopp’s reign.
In part two, I’ll justify our team selection further and run through some players that we have not selected at this stage. My sojourn into blogging continues, although surprisingly no awards just yet.
Behaviour: Vigilant like Becker
We’ll kick things off with the keeper at the back that stops everything – AimBehaviour. Basically your Alisson Becker. Fast, far-reaching, vigilant and watchful, signalling to the rest of the team when a change of security tactic is required. Behavioural biometrics is how you continuously and passively authenticate a user, by the way in which they interact with an application. On a mobile, this data includes pressure, typing speed, angles and more. If it’s a web application, data is analysed from the keystroke and mouse dynamics. In both instances, from the moment an app is opened, data is gathered that creates a profile…and it’s this that’s compared to a user’s template, to confirm it is really them and not an intruder.
Anomaly Detection: Relentless like Milner
AimBehaviour supervises the whole session with a holistic view, constantly shouting advice and sending instructions to teammates. AimBrain’s other module in the behavioural space is AimAnomalyDetection, or in 5-a-side terms, the player that runs & runs & runs. Your Milner, essentially. This player’s fitness level is through the roof, continually working on a particular application to identify any behaviour that is similar to known fraudulent behaviour, so it can be flagged immediately. This module never relaxes, so it’s easy to see why today it’s capturing 75% of our client’s fraud that they could not previously capture.
Facial authentication: Proven like Henderson
Moving through the team sheet, next up is El Capitano, AimFace. Sturdy and dependable. Your Henderson, as it were. Not only has AimFace been a long-standing member of the team and proven it’s worth with multiple clients, but AimFace has demonstrated industry-leading accuracy with an equal-error rate of 0.03% against the independent NIST dataset, FERET. AimFace leads by example, preventing fraud whilst maintaining a great user experience.
Voice authentication – tactical like Van Dijk
The Tactician – your Van Dijk – not only spots and plugs defence vulnerabilities, but brings detailed plans for set-pieces. In our case, this role is taken up by our resident AimVoice. Preventing fraud in multiple channels and languages, AimVoice can be used in an active or passive manner where the module uses a randomised voice challenge for liveliness detection for active authentication, or can passively authenticate an end user when they contact a call centre, for example.
Combination biometrics: unbeatable like Salah
Our last slot is taken by the Clinical Finisher, AimFace//LipSync – your Mo Salah; a relative newcomer that converts any half-decent cross with supreme class. This module marries together facial authentication and a randomised voice challenge, meaning no response can be synthesised in real-time making it our flagship, anti-spoofable module. AimFace//LipSync is a recent signing only six months ago, yet has received a lot of attention through its ability to satisfy the regulator’s requirements around PSD2 and SCA.
So there we have the team. In the second half of the blog, I’ll talk about the team as a whole, and how the skills can be switched up to defend against any attack.