3DiVi speeds up biometric matching, liveness with facial recognition update

3DiVi has rolled out the latest versions of its Omni Platform facial recognition system and BAF, its biometric authentication framework. The updates are promising faster and more efficient performance.

Omni Platform version 1.21.3 delivers a significant performance upgrade for high-load biometric identification systems. The update delivers 33 percent faster face search and a 30 percent boost in profile management operations, according to 3DiVi.

A revamped matcher-service logic drives the 33 percent increase in face identifications per second (RPS) even under heavy, continuous user flows. The company says the enhancement is designed for environments where rapid throughput is essential, such as fast-lane turnstiles and contactless face biometric payments on public transport.

Core profile operations like creating, deleting and linking profiles to watchlists, as well as sample generation now run 30 percent quicker. This performance gain is particularly valuable for deployments managing over 100,000 face profiles, 3DiVi said.

Under the hood, 3DiVi has removed the matcher-router component and streamlined the matcher-service to speed up face comparisons. The team also resolved issues with idle database connections, boosting reliability during peak loads and eliminating the need for manual restarts or interventions.

The new release further reduces resource consumption when processing face images at rates exceeding 100 requests per second, lowering infrastructure costs without sacrificing throughput. An asynchronous cleanup mechanism now clears outdated or unlinked samples, while a smarter index-monitoring system and decoupled index states ensure that large updates in one segment do not interrupt operations elsewhere.

3DiVi has also unveiled BAF version 1.12.0. The biometric authentication framework now delivers three times more efficient single-frame liveness checks alongside enhanced flexibility in WebRTC video streaming. These advances aim to curb false rejections that undermine digital onboarding conversions and frustrate both businesses and end users.

The new selective WebRTC streaming capabilities allow developers to transmit only motion-control video to their servers without coupling it to full liveness-reflection footage. This granular control makes it easier to craft bespoke, multi-step biometric verification workflows, lowers network and server strain, and helps organizations meet rigorous data-storage and processing mandates.

Meanwhile, improvements to the liveness engine enable checks to complete far more rapidly, yielding threefold fewer false positives. The 1.12.0 update is available now. Businesses can request a trial and explore the updated technical documentation on 3DiVi’s website, additionally there is a free online demo.

3DiVi also updated its Face SDK to provide three times better accuracy in biometric liveness detection, adding native support for Go and Node.js last month.

Article Topics

3DiVi  |  biometric liveness detection  |  biometric matching  |  biometrics  |  facial recognition