Download our in-depth report: The Ultimate Guide to IT Security Vendors
Sensory, the Santa Clara, Calif.-based voice, vision, and natural language processing specialist, announced the second major update to its TrulySecure face- and voice-recognition platform this week. In version 3 of the software, the company has enlisted advanced artificial intelligence (AI) to improve accuracy and block crafty attempts to gain access to applications protected by the technology.
TrulySecure is notable for its layered approach to biometrics. Instead of relying on a single identifier like someone's face, Sensory's tech uses both facial and voice recognition to authenticate users. This tactic, combined with the company's AI-derived detection technologies, helps eliminate frustrating retries while enhancing security, said Todd Mozer, CEO of Sensory.
"By layering and fusing face and voice biometrics together we are able to perform more securely through multiple checkpoints, while also enabling more immunity to varying conditions," Mozer told eSecurity Planet. "For example, if a false 'accept' passes one biometric the second biometric could catch it, and likewise an environment that might cause a false 'reject,' like extreme noise for voice verification, could be accepted through the face authentication."
The latest version of TrulySecure uses upgraded deep learning algorithms to improve detection, even when part of a person's face is partially obscured by glasses or an accessory on the face. In this latest iteration, the company estimates that the solution's accuracy has been improved by more than 50 percent and the ability to fool it using photos or video footage has been further reduced using the product's improved "liveness" detection.https://o1.qnsr.com/log/p.gif?;n=203;c=204650394;s=9477;x=7936;f=201801171506010;u=j;z=TIMESTAMP;a=20392931;e=i
According to the company, TrulySecure boasts a 99.999-percent accuracy rate without requiring costly biometric scanners. And when TrulySecure is used as a login tool, users don't have long to wait for access to their apps.
"The login process occurs quickly and seamlessly," Mozer assured. "In a voice activated assistant the mere act of calling the assistant could serve the dual purpose of identifying the user, likewise the person's face can be identified immediately with no special touching or user manipulation."
Those capabilities are now for iOS and Android mobile devices, in the form of SDKs (software development kits) for developers wishing to integrate the security-enhancing biometric technologies into their apps and solutions that company to FIDO (Fast IDentity Online) specifications. Android users can check out Sensory's tech by downloading the AppLock by Sensory app from the Google Play app store.