Oosto offers guidance on ethical face biometrics deployments, platform upgraded

Oosto has published a white paper on ethical facial recognition deployment and announced a series of next-generation capabilities added to its computer vision platform.

‘The Rise of Ethical Facial Recognition’ is a guide for customers to help them implement ethical face biometrics. As such the 24-page white paper discusses the technology’s benefits and provides basic information on use cases and how facial recognition and verification work.

The company contrasts law enforcement and commercial applications of facial recognition, in terms of how they are used and the sizes of databases involved, and provides a primer on biometrics terminology.

An exploration of the ethical challenges of facial recognition identifies insufficient training data, data sources, the use of the Fitzpatrick Skin Type scale for skin tones, reference data quality and user consent as potential problem areas. The quality of researchers and need for operational due diligence are also covered.

Oosto urges employee training, human oversight, customizable privacy protections and limited system access as best practices for ethical facial recognition.

The company explains its biometric data storage policies, and touches on the source of its training database of 50 million images, which includes video sources.

Three tips for optimal watchlist database creation are offered, and setting confidence thresholds is discussed.

New computer vision features

Oosto has also expanded alerts on its OnWatch security platform, previously known as A Better Tomorrow, beyond biometric watchlist matches to notify security operators when any known or unknown individual approaches a restricted space.

The set of upgrades also includes improvements to the performance of its mask detection, recognition of masked faces, and alerting. Oosto’s video forensics capabilities now include fast offline searches, with the company saying face biometric searches for a person of interest can cover hours of video within seconds.

Other improvements include forensic cross-referencing, more complete audit logs and integrated account management functionality.

“We are committed to delivering the best performance and ease-of-use for our customers,” says Ido Amidi, VP of Product Management at Oosto. “By integrating Vision AI, we’re able to help our customers better protect their premises while enhancing employees and visitors’ experiences.”

Article Topics

biometric identification  |  biometrics  |  commercial applications  |  computer vision  |  ethics  |  face biometrics  |  facial recognition  |  law enforcement  |  mask detection  |  Oosto  |  video surveillance