Partner-ready Generation 5 release improves accuracy, speed
and deployment flexibility, adds new mobile capabilities

SAN FRANCISCO – Paravision, the U.S.-based leader in mission critical computer vision, announced today the availability of its fifth-generation face recognition product suite, demonstrating substantial improvements in accuracy, deployment flexibility, and speed. The newly released technology is the same as the NIST benchmarked in its most recent FRVT 1:N test, delivering a 70% decrease in error rates and ranking Paravision as the #1 face recognition vendor for accuracy worldwide. 

Paravision also announced new face recognition SDKs for mobile devices running both Android and iOS, as well as updates to the company’s Edge AI SDK for Ambarella. The new capabilities enable Paravision’s partners to deploy face recognition in a more flexible way and optimize for speed, accuracy, scalability, or cost.


Our fifth-generation products mark significant milestones in both the accuracy and usability of face recognition. Accuracy is essential because reducing errors improves inclusivity and performance for everyone. Usability is fundamental since face recognition must work in a wide variety of environments and adverse conditions. Equally important to us is that face recognition must be ethically trained and conscientiously deployed. We’re proud that Paravision has not just the most accurate face recognition but we’ve achieved it with AI ethics at the forefront.

Joey Pritikin, Chief Product Officer at Paravision

Paravision has consistently been recognized as a top performer in NIST’s Face Recognition Vendor Test (FRVT) reports, and was recently ranked as the most accurate face recognition provider in the world in NIST’s 1:N test in January 2022.

Unlike many areas of Artificial Intelligence, the performance of face recognition algorithms is independently tested by NIST. Paravision is the only U.S.-based company to consistently rank among the top in its tests, meaning partners can be confident in the performance and integrity of our platform.
Benji Hutchinson, COO at Paravision

The fifth generation also includes a new release of Paravision’s liveness and anti-spoofing toolset for Ambarella-based Edge devices. This now enables AI-based prevention against a wide range of spoofs and presentation attacks from paper printouts to high definition videos to advanced 3D masks made of paper, resin, or latex.

As with prior releases, Paravision’s fifth-generation face recognition is available as both SDKs and cloud-ready, Docker-based matching engines, optimized for leading compute platforms and operating systems from the server-side to the extreme network edge, with toolsets supporting major programming languages.  Paravision’s new mobile products are based on best-in-class face recognition technology and enable modular deployment to help partners optimize user experience, latency, and security.

Paravision’s Gen 5 face recognition product suite is available now for all Paravision partners worldwide.