A new biometric face-matching platform for enterprises has been launched to commercial availability by Paravision to meet the needs of demanding use cases.
Paravision Search was first announced at Converge 2022, and is based on the company’s fifth generation facial recognition model, which has been evaluated in NIST’s FRVT. The company claims the model ranked as the most accurate in the FRVT in a January update of the ongoing NIST test just over a year ago.
The company’s facial recognition was also among the top performing matching systems in the latest DHS biometrics rally.
Paravision says the new platform can enroll 50 million identities in less than 3.5 hours and perform 1.4 million matches per second per core when running on an Intel Xeon processor at 3.1GHz. The platform can handle 120 concurrent requests per second on 16 CPUs against 100,000 templates, and can be further scaled with additional hardware.
Search also features dynamic sub-gallery creation and programmable elasticity for rapid re-allocation of computing resources to meet demand spikes and reduce costs when traffic is lower. The gallery management tool functions as a distributed database, Paravision says, allowing data to be added or searched quickly, with sub-galleries and attribute-based filters.
“Paravision Search offers customers unparalleled performance and scalability in an easy-to-deploy package that’s typically ready to use within a week,” says Paravision Chief Product Officer Joey Pritikin. “It is a result of our continued commitment to providing state-of-the-art solutions that cater to the diversity of use cases we serve. Paravision Search will help our partners enhance their security posture and deliver superior experiences while minimizing development risk and driving significant cost savings.”
The platform is designed for an easy and flexible deployment process, according to the announcement, with architecture suited to multiple use cases and modular features. It is optimized for large-scale ID verification applications, but also for smaller, more dynamic event-day applications.
“We’re proud to integrate our NIST-ranked face recognition technology into a truly enterprise-grade, feature-rich search system,” comments CTO Charlie Rice. “Coupled with its native elasticity and cloud-ready architecture, Paravision Search allows organizations to scale with high confidence, making it the ultimate matching platform for high-volume use cases.” A new biometric face-matching platform for enterprises has been launched to commercial availability by Paravision to meet the needs of demanding use cases.
Paravision Search was first announced at Converge 2022, and is based on the company’s fifth generation facial recognition model, which has been evaluated in NIST’s FRVT. The company claims the model ranked as the most accurate in the FRVT in a January update of the ongoing NIST test just over a year ago.
The company’s facial recognition was also among the top performing matching systems in the latest DHS biometrics rally.
Paravision says the new platform can enroll 50 million identities in less than 3.5 hours and perform 1.4 million matches per second per core when running on an Intel Xeon processor at 3.1GHz. The platform can handle 120 concurrent requests per second on 16 CPUs against 100,000 templates, and can be further scaled with additional hardware.
Search also features dynamic sub-gallery creation and programmable elasticity for rapid re-allocation of computing resources to meet demand spikes and reduce costs when traffic is lower. The gallery management tool functions as a distributed database, Paravision says, allowing data to be added or searched quickly, with sub-galleries and attribute-based filters.
“Paravision Search offers customers unparalleled performance and scalability in an easy-to-deploy package that’s typically ready to use within a week,” says Paravision Chief Product Officer Joey Pritikin. “It is a result of our continued commitment to providing state-of-the-art solutions that cater to the diversity of use cases we serve. Paravision Search will help our partners enhance their security posture and deliver superior experiences while minimizing development risk and driving significant cost savings.”
The platform is designed for an easy and flexible deployment process, according to the announcement, with architecture suited to multiple use cases and modular features. It is optimized for large-scale ID verification applications, but also for smaller, more dynamic event-day applications.
“We’re proud to integrate our NIST-ranked face recognition technology into a truly enterprise-grade, feature-rich search system,” comments CTO Charlie Rice. “Coupled with its native elasticity and cloud-ready architecture, Paravision Search allows organizations to scale with high confidence, making it the ultimate matching platform for high-volume use cases.” Read More
