Omnigarde LLC has announced that its latest AI-powered face recognition software, Omnigarde-004, has achieved a top three position among U.S. companies and a top 15 position globally in the National Institute of Standards and Technology‘s (NIST) Face Recognition Vendor Test (FRVT) 1:1 verification test.
The Omnigarde-004 algorithm has shown improvements, reducing the average error rate by 25 percent compared to its previous version and by over 90 percent compared to the first version released in 2021.
The NIST FRVT 1:1 verification test evaluates face recognition software under various conditions, including matching mugshot-type images, border-taken images, and kiosk-taken images. Omnigarde-004 performed well across these scenarios, demonstrating its capability to handle different image types and conditions. The test also measures factors such as memory usage, feature extraction time, and matching time, providing a comprehensive assessment of each software’s performance.
Omnigarde, founded in March 2021, focuses on developing advanced biometric technologies, including face, fingerprint, and voice recognition using AI. The company aims to be a leading provider of biometric technologies and has achieved notable performances in various industry benchmarks.
“Omnigarde-004’s success underscores our commitment to innovation and excellence in the field of biometric identification,” said Dr. Peter Lo, Founder and CEO of Omnigarde. “As the competition in the NIST FRVT 1:1 test intensifies, Omnigarde remains at the forefront of cutting-edge technology, driving advancements in face recognition software.”
The NIST FRVT provides an independent evaluation of face recognition software, offering insights for government agencies, industry stakeholders, and the research community, while the test’s ongoing nature allows companies to refine their technologies and submit updated versions for evaluation.
This week, NIST released its latest Face Recognition Technology Evaluation (FRTE) report, showcasing advancements in face recognition algorithms from various vendors. The report highlights significant accuracy gains attributed to deep convolutional neural networks (CNNs), which have improved the technology’s ability to handle low-quality images and pose variations.
–
July 11, 2024 – by Tony Bitzionis
