Sensory releases major update to its TrulyHandsfree speech recognition platform
August 6, 2015 –
Sensory Inc. announced it has released TrulyHandsfree 4.0, the latest version of the company’s embedded small-footprint voice user interface platform.
TrulyHandsfree 4.0 offers improved performance, up to an 80% increase in word accuracy, filterbank features as an alternative to MFCC features, and new phrase spotting techniques.
The new release also includes Sensory’s new neural nets engine that supports deep learning acoustic models, which can significantly improve speech recognition accuracy in real-world noise.
The platform was designed to take benefit from deep learning and other advanced techniques to improve its overall performance and accuracy.
Sensory uses a unique form of a neural network with deep learning to achieve acoustic models significantly smaller than the present state-of-the-art networks, while employing advanced speech feature extraction to generate high accuracy in noisy environments.
The new version features new spotting techniques to ensure that the core part of the user’s spoken request can be recognized in the middle of speech, or when hindered by ambient noise.
The technology is also able to fully operate with less than a 2mA draw on the battery, and as low as 1mA with Sensory’s low power sound detection intellectual property.
TrulyHandsfree 4.0 offers several improvements over the previous version, including larger deep learning acoustic models with higher accuracy, new algorithms that overcome reverb and harsh acoustic environments, advanced filterbank features that improve front-end speech feature extraction, full compatibility with TrulyNatural, and improved architectural scalability which allows for low-power DSP implementations with secondary accuracy improvements at the operating system level.
“TrulyHandsfree was an industry-first that literally changed the way people interact with devices of all kinds,” said Todd Mozer, CEO of Sensory Inc. “Many thought it was impossible to create an always-listening voice user interface, that had both accuracy and low power, but Sensory did it, and the industry has followed but trailed in performance. TrulyHandsfree 4.0 takes performance to a whole new level with an accuracy, footprint and power consumption that others just can’t touch.”
TrulyHandsfree 4.0 supports US English, UK English, French, German, Italian, Japanese, Korean, Mandarin Chinese, Portuguese, Russian, and Spanish, while TrulyHandsfree SDK is available for Android, iOS, Linux, QNX and Windows.
Sensory is also offering ultra-low-power deeply embedded ports of TrulyHandsfree for DSP/MCU IP cores, as well as for integrated circuits.
Previously reported, Sensory Inc. released an a la carte SDK for the new embedded speaker verification component of its TrulySecure technology platform.
Leave a Comment
Via: Google Alert for Deep Learning