Face recognition systems are getting hacked: it’s time for liveness checks
Performing Know Your Customer (KYC) checks is the essential first-step in on-boarding customers. KYC process has travelled a long way from being paper-based and data-centric to technology-powered image / video-based validation of a customer’s identity. This has led to a shift in the paradigm from merely being able to verify the identity of a customer to validating if they are indeed the person they claim to be.
Facial Recognition (FR) is a technique for identifying and validating an individual’s identity using their facial features. FR is gaining traction due to increased digital penetration and is touted to be one of the most reliable and secure means of authentication. However, with fraudsters and scam artists upping their game, there has been an alarming increase in identity thefts and cyber frauds.
Fraudsters use spoofing techniques that exploit the vulnerabilities in bio-metrics-based FR systems to commit acts of impersonation and crime. Some of the commonly used spoofing techniques include high-quality silicon masks, synthetic video sequence created out of a stream of pictures and hi-definition video replays. It hence becomes imperative for businesses to stay ahead of the curve by adopting techniques like Liveness Checks which has become the mainstay security technology for facial biometric validation.
Liveness check, used in tandem with facial recognition, entails determining if the image / video captured during the KYC process belongs to a living subject present in the verification session. Built on AI and Deep Learning technologies and armed with a strong repertoire of facial recognition algorithms, these checks help ward off impersonation attempts. There are two major categories of liveness detection checks – active and passive.
- Active liveness detection requires the user to perform some actions / gestures during the session. These could include speaking out based on textual / image-based cues or performing body gestures in response to screen commands. Active mode improves security and accuracy of the verification process.
- Passive liveness detection does not involve any action or gesturing from the user, but just warrants a selfie-video of the user, running for a specific duration. The video / image thus captured are compared against the reference picture, in addition to being analysed by advanced Deep Learning (DL) models for distortions and manipulations. It is simple and quick, without any exertion on the customer.
With proliferation of online frauds and sophisticated impersonation techniques employed by cyber scammers, Facial Recognition technology needs to be adequately supplemented by liveness checks. Digid provides businesses with a robust technology suite that enables interactive, secure and customer verification. Liveness check is seamlessly embedded in the automatic KYC process, providing a delightful customer experience.
How Digid performs liveness checks?
As part of the approval process, done on the business side, the following validations are done by the approver. These validations entail both active and passive liveness checks.
- Verification of image consistency between the documents uploaded and the selfie-video
- Matching the OTP gestured by the customer and the actual OTP
- Signature checks by comparing signature in the uploaded identity documents against the signature captured during the video session
- Spoof detection checks
Digid ensures fool-proof liveness checks on the selfie-videos that are captured by the customers at the time of on-boarding, thereby preventing impersonation, frauds, business losses and enforcing compliance. A win-win proposition for both businesses and their customers – Digid is the way to go!