San Francisco – January 14, 2025 – Paravision, a leader in trusted Identity AI, today announced that its upcoming 7th Generation Face Recognition has ranked in the global top 5 in the primary benchmark of the latest NIST Face Recognition Technology Evaluation (FRTE) 1:N test. The company also maintained its distinction of being the #1 ranked, most accurate vendor from the U.S. and Europe.

Beyond its leading performance in the key “visa-border” test scenario, Paravision achieved a top 10 ranking across all tested scenarios, a feat achieved by only five vendors globally. This includes performance across diverse test cases, showcasing the consistency and reliability of Paravision’s technology.

 

Gallery Probe Gallery Size Paravision Global Rank
Mugshot Mugshot 12M 8
Mugshot Mugshot 1.6M 8
Mugshot Webcam 1.6M 9
Mugshot Profile 90° 1.6M 8
Visa Border 1.6M 5
Visa Kiosk 1.6M 4
Border Border ΔT≥10 YRS 1.6M 10
Mugshot Mugshot ΔT≥12 YRS 3M 7

Paravision is also one of only three vendors worldwide with four separate submissions ranked in the top 30 on NIST’s 1:N Identification by Algorithm leaderboard. These submissions include Paravision Gen 5, Gen 6, and Gen 7 Face Recognition models, underscoring the enduring quality of the company’s technology and its ongoing commitment to innovation and continuous improvement.

In its latest submission, Paravision Face Recognition showcases significant advancements over its predecessors, including:

  • 42% improvement in error rates compared to Gen 5
  • 15% improvement compared to Gen 6

Our upcoming Gen 7 technology reflects years of relentless focus on performance, accuracy, and innovation. Being consistently ranked as a leading vendor across scenarios and generations–and as the highest-ranked U.S. and European provider–underscores our ability to deliver technology that excels in both the present and future.
Charlie Rice, Paravision CTO

The National Institute of Standards and Technology (NIST) Face Recognition Technology Evaluation (FRTE) is the gold standard for evaluating the performance of face recognition technology. Its independent and rigorous testing ensures that vendors are objectively assessed across accuracy and reliability metrics, making NIST results the definitive benchmark for the industry. The NIST FRTE 1:N (one-to-many) test specifically evaluates how well algorithms can accurately identify a person from a database of millions of individuals. This scenario closely mirrors real-world applications like identity verification, border security, biometric payments, and access control, where the technology must perform reliably under diverse conditions. High rankings in this test demonstrate the robustness, scalability, and practical relevance of a vendor’s face recognition technology.

We’re proud to have our production technology ranked so high in NIST FRTE. While NIST rankings are the gold standard of face recognition benchmarking, what really sets us apart is our consistent performance across a wide range of third-party evaluations, including the DHS S&T Biometric Technology Rally and other real-world assessments. These results demonstrate not just our strength in controlled benchmarks, but also our ability to deliver reliable, high-performing solutions in diverse and demanding environments.
Joey Pritikin, Paravision Chief Product Officer

The Gen 7 model will officially launch later this year, redefining industry standards for accuracy and dependability in applications ranging from border control to identity verification. Gen 7 will also be fully integrated into Paravision Search, enabling high performance, cost-effective deployment in large-scale identity verification systems, as rigorously tested by Identity E2E last year.