Indian startup HyperVerge bags two global recognitions

The first achievement is for their advanced facial recognition technology, which positions the company as the only Indian company to be in the top 15 per cent of the global leader board for the National Institute of Standards and Technology's (NIST) Face Recognition Vendor Test (FRVT).

Additionally, they became one of the three organizations worldwide to receive the global ISO certification (ISO-30107-3) for single-image liveness detection solutions, said the enterprise.

NIST FRVT is a global leaderboard for rating the effectiveness of facial recognition systems for enterprise, civil, law enforcement, government, and security applications based on accuracy, speed, storage, and memory criteria.

Elated with the achievement, Vignesh Krishnakumar, Chief Technology Officer, HyperVerge, said, “We are extremely delighted to have been recognized on globally accepted standards for Face-Match and Liveness detection technologies. This further strengthens HyperVerge’s position as the preferred partner for AI-enabled onboarding and identity verification use-cases across the world. This is aligned with our mission to make KYC and identity verification processes seamless, completely automated, and fraud-proof while being fully compliant”

Further elaborating, Vignesh said, “To simplify, let’s look at an example. Ramnath, a 91-year-old Army veteran from Uttarakhand, gets a monthly pension from the Govt of India. Every year, he has to travel 100 km to visit the nearest Pension office to prove he is alive & get a life certificate. He can now generate the life certificate comfortably sitting at home using HyperVerge powered Face-match & Liveness solution.”

HyperVerge’s single image liveness detection solution was put through rigorous testing with attempts to break the system for eight hours with digital screens, posters, masks and other devices, with zero successful break-ins by the ISO-30107-3 evaluators.

The ISO-30107 standard certifies the secure and effective performance of the feature under review.

Their face-recognition submission was evaluated on accuracy, efficacy, and security on a variety of datasets, and positioned 33rd on the list which contains 331 entries.