Jumio develops in-house biometric liveness detection capability

Jumio has launched a biometric liveness detection capability developed in-house to defend against advanced fraud, including injection attacks and deepfakes.

The new Jumio Liveness combines presentation attack detection (PAD) with AI models that defend against bypass attacks like video injections. It checks if multiple people are present during the selfie capture process, detects coerced verifications and if people are unconscious or sleeping, according to the company announcement.

Injection attacks have become recognized as a significant threat vector for organizations, with the Kantara Initiative recently launching a dedicated discussion group to research what can be done about them.

Jumio CPO and CTO Bala Kumar says the technology aligns with the company’s goal of protecting against fraud without adding complexity to the user journey of its clients’ legitimate customers.

The company emphasizes its layered approach to fraud protection, and the backing of its liveness detection and anti-spoofing technologies by an extensive intellectual property portfolio that spans nearly 100 patent families. These measures are combined with Jumio’s anomaly and suspicious activity monitoring, which provides a real-time feedback loop that allows the technology to quickly adapt to threats.

Liveness detection and downstream processes like biometric authentication are also made more reliable by Jumio’s automated capture technology, the company says. The feature guides users through the capture process with instant feedback to ensure optimal image quality with clear, glare and blur-free selfies. This also helps keep conversion rates high for valid users.

Jumio has previously used liveness detection capabilities provided by third parties, such as iProov, from which Jumio integrated PAD with its KYX Platform in 2021.

The new Liveness software was introduced at Money20/20 USA in Las Vegas on Monday.

Article Topics

biometric liveness detection  |  biometrics  |  injection attacks  |  Jumio  |  Jumio Liveness  |  presentation attack detection  |  research and development