Sensory, Inc. has released a new SDK that allows devices and applications to dynamically switch user profiles via voice recognition. Called TrulySecure, the SDK is available for device integration as either text dependent or text independent.
The technology supports multiple user profiles and is able to distinguish between each associated voice; and the more it’s used, the more accurate it gets. Sensory says that it has a success rate of up to 96 percent in accurately identifying users, and a 99.99 percent success rate at identifying impostors.
Speaking in a press release, Sensory CEO Todd Mozer explained that the SDK’s development is the result of strong interest from clients, and emphasized its versatility. “What sets this technology apart, other than its incredible accuracy,” he said, “is the fact that it can be incorporated into devices in either text-dependent or text-independent modes.” He elaborated that “a passphrase may be pre-determined, or left to the end user to choose, or the application may identify the user from their unique characteristics from random speech.”
The TrulySecure SDK follows closely after Sensory’s release of its TrulyNatural SDK in May, and complements the company’s low-power voice recognition technology, developed in collaboration with software firm Synopsis Inc., unveiled back in March. That technology was designed specifically for the Internet of Things, and while the TrulySecure SDK has any number of potential applications, it’s quite possible that it could find its way into some IoT devices as well, distinguishing between various users in smart home or business settings, for example. Given the SDK’s versatility, this and many other potential applications are possible.