Sensory Releases Embedded AI That Allows Voice Assistants and IoT Devices to Identify a Wide Variety of Home Sounds
Sensory, a Silicon Valley-based company focused on improving the user experience and security of consumer electronics through embedded AI technologies, announced TrulySecure Sound ID, a development for cloud-free, always-listening AI. It gives devices the ability to identify a variety of common and critical sounds within the home, and intelligently interpret if action needs to be taken, without the security risk of sending audio recordings to the cloud for processing.
Backed by Sensory’s 25 years of experience developing practical applications for neural networks and deep learning, TrulySecure Sound ID is capable of recognizing a variety of environmental sounds, including glass breaking, babies crying, dogs barking, home security alarms, smoke/CO alarms and low-battery warnings, doorbells, knocking, snoring, and more. Consumer devices could instantly warn users when these sounds occur and send the owner sound clips to better understand the situation.
TrulySecure Sound ID says it boasts performance and privacy by performing all analyses on device. By training each specific sound profile on thousands of real-world samples, similar to how Sensory trains biometric recognition, TrulySecure Sound ID takes several discriminate factors into account when processing sounds. Sound ID learns each sound type utilizing an approach that combines deep learning model training with proprietary discriminant learning techniques, the two approaches combined produce a superior solution than either approach could achieve in isolation.
Sound ID is available today as a component of Sensory’s TrulySecure Speaker Verification (TSSV) suite of AI solutions for smart speakers and IoT devices. Ideal for enabling smart assistants, TSSV combines:
- Text dependent and text independent voice biometric recognition for user ID and authentication
- Always listening wake word recognition
- Seamless voice enrollment
- Sound ID
The company says all of the technologies in TSSV pose no privacy or security concerns because they run completely on device and never touch the cloud. TSSV supports all major operating systems and is hardware agnostic, offering nearly limitless implementation flexibility. Additionally, Sensory can customize TSSV to match the exact needs of its customers, enabling only the sound profiles required for specific use cases.
Voice biometrics is becoming popular with big corporations but is still out of reach for many
Designed to not invade user privacy or record user conversations, Sensory's VoiceGenie Smart Speaker reference design is a low-cost solution for enabling voice-powered music control on Bluetooth speakers.
The IVA market is anticipated to grow significantly over the forecast timeframe due to the growing demand for online self-service, self-reliance, and rapid query resolution requirements. Websites are becoming the preferred service channels to gain information, make purchase decisions, and resolve issues.