1. Why do onboarding dropouts happen?
A dropout can occur when your customers fail to verify their identity during the onboarding process. This is important to note, because some vendor’s liveness solutions are not optimized for uniform levels of accuracy.
Common reasons for dropouts:
- Your customers give up on signing up to your service because it becomes too complicated
- Your customers fatigue when they need to repeat actions in front of a camera
- Your customers cannot be verified because the solution has limited technical capabilities.
- Yours customers don’t feel comfortable showing their face
We offer a simpler technical solution to the sign-up process.. Our identity verification platform combines best-in-class passive liveness detection with facial verification and identity document verification.
We offer a variety of flexible ID verification products to suit companies of all types and sizes, across different industries and use-cases, globally.
2. How do biometric liveness detection and facial authentication techniques work?
Various buzzwords are used throughout the global identity verification management industry. Some of these might seem a little confusing at first, given how some vendors will eagerly seek to demonstrate their claim to have the latest and greatest technologies, making it less clear who to actually trust.
But we’re here to help you understand exactly the benefits of liveness detection and demystify the onboarding process, breaking down some of these terms to help you understand the customer onboarding process better.
According to the Encyclopedia of Biometrics, Liveness detection reduces the risk of spoofing (the process by which fraud occurs when somebody tries to pretend they are somebody else) by requiring a liveness signature in addition to matched biometric information.
There are two main different types of biometric liveness detection methods used in the industry:
- Active liveness
- Passive Liveness
Both are commonly known as facial verification, not to be confused with facial recognition. The difference is important because facial recognition refers to two or more images compared against each other. You would usually find facial recognition systems in public places where images captures are matched against a central database (e.g airport surveillance systems).
3. Why are active liveness detection models flawed?
The main idea behind active liveness detection models is to ‘actively’ challenge the user to provide a predictable response for the computer to understand.
For example, a customer might be prompted to move their head in one direction, blink in another or even smile or frown on command. This is a poor alternative to passive liveness, as these critical facial authentication captures leave themselves wide open to fraudulent exploits.
Active liveness models also fatigue the end user and significantly increase the risk of failing the verification process entirely with more drop-outs. The more drop-outs you have, the more unhappy customers you risk losing.
Is that a risk your business can afford?
4. What is passive liveness detection?
A passive liveness model is a frictionless user-experience. It does not require the use of tiring and tedious action commands. Instead, passive liveness is powered by a sophisticated software engine built on machine learning and computer vision technologies.
Passive liveness works seamlessly in the background by ensuring the customer’s face can be verified and authenticated accurately without further user input. Our passive liveness product only requires a few seconds of video to verify an identity accurately. And customers can still take a selfie video under a variety of challenging lightning environments, which leads to fewer dropouts.
5. What biometric liveness detection method is best?
During the digital identity verification process, a liveness detection checks is necessary to determine that the customer is who they say they are. During the onboarding process, a customer will usually be instructed to take a ‘selfie’ video with their smartphone to establish the ‘liveness’ of the customer vs someone trying to spoof identity.
Active and Passive liveness modes are known for their distinct technological differences, including accuracy and speed methodologies. It is important to recognise that liveness checks need to be tailored to business and data-sensitive industries (telecommunications, banking, fintech), which are more likely to require a technology solution that is less prone to error and highly secure.
Passive liveness detection is considered to be the most accurate identity verification tool on the market today. But is not offered by many vendors due to the complexity of the technology and the ongoing cost to research to ensure this tech is continually updated to reflect the latest security concerns.
6. Why is passive liveness detection considered industry best practice?
Anti-spoofing algorithms contained with our biometric passive detection engine make it increasingly difficult for fraudsters to ‘game the system’ using 3D-scanned masks, deep fake videos and other genuine presentation attacks to fake identity.
7. What is active liveness detection?
Today, active liveness is the most commonly used form of presentation attack detection (PAD) in the industry.
In fact, more than 99% of companies in the ID verification industry still use this method to verify their customer’s identity. Why? Because companies are lacking the expertise to build complex machine learning and computer vision algorithms and the technology is relatively new.
There aren’t many vendors of Passive Liveness Detections offering an accurate and mature product yet.
8. How do you verify identity without a human? Can I trust a computer?
Some customers might be surprised to learn that we do not use a human to verify identity, a technique commonly known as ID verification over video.
Our deep learning networks have been tested over many years in collaboration with our academic partnership at the ETH University in Zurich, Switzerland.
Our research has taught us how to analyse faces from different backgrounds, different ages and different sexes. Every discernible facial difference makes a difference to how machines or humans interpret faces.
9. Facial verification: Are humans or machines more effective at identity verification?
People are prone to fatigue and tiredness. They make mistakes and they can be biased to different face types without realizing it. (link)
An algorithmic A.I model shares none of these limitations.
Instead, computer models have been rigorously tested to look for 1000s of variations in human facial characteristics and expressions under challenging lighting conditions, using different camera lens apertures across a vast array of smartphone devices.
When you consider all the different ways to take a video selfie, how you could trust anything other than a passive liveness model?
10. How do presentation attacks work?
We’ve already spoken a little about fraud and the inherent dangers associated with global risk and poor technological solutions. Presentation attacks describe the process of trying to fraudulently ‘spoof’ identity systems with false credentials. This is a major risk for businesses during the digital onboarding process.
While your customers might not be familiar with the latest attack modules, there are different types of presentation attacks you should be aware of:
- 2D attacks. Using printed images, attackers will seek to deceive the computer, with an image-based attempt to verify identity.
- 3D attacks. The most advanced use high quality 3D-printed masks based on the requirements of the user’s identity being spoofed
These attack types usually align with the intended results and the goals of the fraudster.
Passive liveness provides your business with confidence that your customers are who they say they are.
11. Integration Options: On-premises vs. Cloud
We offer all customers two main integration solutions:
- Off-Premise (cloud)
Our identity verification platform is designed to be integrated to work seamlessly on your servers (on-premise) or hosted on ours (cloud).
Both integrations allow for exactly the same functionality.
However, not all vendors offer this functionality in the same way. When we say on-premise integration, we genuinely mean it.