Paravision: Paravision builds on RIVTD success with Liveness 2.0 release

Paravision will not rest on its liveness laurels. Fresh off a successful appearance by its presentation attack detection (PAD) system in a major industry evaluation, the company is launching Paravision Liveness 2.0, which significantly improves on its already-impressive accuracy and speed, Paravision CPO Joey Pritikin told Biometric Update in an email interview. The company's liveness detection solution was one among 15 selected to represent the state of the art by the U.S. Department of Homeland Security's Science and Technology Directorate (S&T). Paravision revealed that it was vendor "P8" in the anonymized Remote Identity Validation Technology Demonstration (RIVTD) Track 3: Liveness Detection to celebrate a top result in combined error rate. The 0.3 percent bona fide presentation classification error rate (BPCER) and 3.3 percent attack presentation classification error rate (APCER) each placed among the leaders in the RIVTD passive PAD class. Further, the combined error rate of 3.6 percent was far lower than the second (8.5 percent) and third (16 percent) lowest rates.Passive convenient"When we think about why passive liveness matters, it really comes down to the user experience," says Pritikin when asked why Paravision decided to develop passive PAD technology. "Passive liveness runs seamlessly in the background - no blinking, head movements, or extra steps - which makes it faster, more natural, and much more scalable across real-world use cases." "If you look at industries like financial services, travel, and border control, security is obviously a priority, but speed and convenience are also critically important. People don't want to be slowed down by a clunky authentication process, and businesses don't want to introduce unnecessary friction for their users, which can lead to losing customers during onboarding or over-reliance on slow, manual fallbacks. That's why we believe passive liveness is the future." Paravision's liveness detection was one of four found to be the fastest among all passive and active liveness solutions tested in the RIVTD. Pritikin acknowledges the debate within the industry about whether passive PAD can match the accuracy of active approaches, but notes that the RIVTD showed passive PAD systems can be just as secure, "while offering a far superior user experience." Paravision Liveness 2.0 is delivering error rates more than 90 percent lower than the version in RIVTD, with the same frictionless experience, he says. Like facial recognition compared to other modalities, passive liveness has caught up.Insights from RIVTDThe RIVTD assessment is a significant moment in the maturation of the field of biometrics liveness detection, Pritikin argues. "It's the first time there has been a rigorous, open, multi-vendor operational test of presentation attack detection technologies to include both passive and active PAD approaches. So in many ways we're getting our first real insights into how different vendors and different approaches stack up." He expresses pride in the company's performance in RIVTD, but says "we also are looking objectively at our 3.3 percent APCER (a measure of false positives), which didn't pass DHS's 3 percent goal, and was not at 0 percent, which may be a requirement for some use cases." The company drew multiple insights from the results. "First, it is clear that we recommended a threshold which may have over-indexed on usability, and a different threshold could have delivered enhanced security," Pritikin says. "Threshold setting is frequently an interactive effort with our partners after preliminary testing in their use case. Second, we're really excited about what performance looks like when optimizing thresholds in combination with Paravision Liveness 2.0, which decreases BPCER (a measure of false negatives) by more than 10X." Liveness 2.0 he says, "delivers dramatically better performance than what we showed in RIVTD, with a goal of delivering best in class APCER and BPCER." The company's commitment to both third-party testing and continuous development were significant contributing factors in its selection as a liveness detection Pioneer in the 2025 Face Liveness Market Report and Buyers Guide from Biometric Update and Goode Intelligence.Ongoing improvementParavision Liveness is trained and tested across a wide range of hardware, Pritikin says, to support various deployment options. Low-quality desktop webcams have been a particular pain point for PAD development, he notes. "Developing a strong PAD solution isn't just about accuracy - it's about ensuring that accuracy holds up across different devices, lighting conditions, and camera qualities. A front-facing camera on a smartphone is very different from a fixed-position camera on a kiosk, and you shouldn't extrapolate performance from one to the other." The company also offers an optional Validity SDK to give users real-time feedback on factors like lighting, face positioning, and image clarity to ensure a high-quality face capture. Paravision also offers a Deepfake Detection product which is separate from Paravision Liveness, as they "are two very different challenges." "Notably, we're one of the only Identity AI providers offering detection of both deepfakes and liveness attacks," Pritikin says. "That means businesses using Paravision can secure both physical and digital attacks - a critical advantage as fraud methods become more sophisticated." The launch of Liveness 2.0 improves the software for customers in industries where security matters most, according to Pritikin. "When you look at sectors like payments and banking, travel, and government programs, the stakes are incredibly high," he points out. "Fraud prevention needs to be rock-solid, but the user experience also needs to be seamless. That's why we focused on making Paravision Liveness 2.0 the most secure, accurate, and frictionless version yet. "The DHS RIVTD test confirmed that Paravision's passive liveness was best-in-class with our previous release, but Paravision Liveness 2.0 pushes the bar even higher by decreasing error rates by over 90 percent compared to the RIVTD-tested version - making it the best solution for high-assurance identity verification in mission-critical use cases." Pritikin advises businesses looking into upgrading their biometric authentication stack in 2025 to consider that security and usability must go hand-in-hand, with the DHS RIVTD test making the importance of speed clear. Fraud techniques like deepfakes and synthetic identity are evolving quickly, necessitating a layered approach to spoof attack defense. And in another insight from RIVTD, usability is a critical factor in mitigating bias, especially around the age of users. This is a sponsored post. For information about advertising, please contact us.