Santa Clara, CA
Device and equipment control: In-home device control (DVC), Toys and Games (T&G)
Core technologies: Speech recognition (SR), Speaker Recognition (SKR)
Scope: Specialized (SPC)
- Sensory focuses on speech recognition technology that can be run on the edge, e.g., embedded in a smartphone or home speaker.
- A specialty is always-listening embedded speech recognition used as a wake-word or wake-phrase, that is, to alert a device it is being addressed so that it doesn’t process arbitrary speech not addressed to the device.
- Sensory’s embedded speech recognition can be expanded into limited command words or phrases or to recognize a list of keywords within an utterance.
- The limited speech recognition can be expanded to more complex vocabularies if processing power on the device allows.
- The company’s biometric speaker identification can be used for authentication.
Sensory licenses embedded AI to differentiate products and make them safer and easier to use. These include wake word, small-to-large vocabulary speech recognition, and natural language understanding technologies. Its biometric recognition technologies are used to unlock a device or authenticate users.
Sensory’s wake word and phrase spotting technology is designed for fast response, low power consumption, and to work from a distance or in noisy environments. Speech following the wake word can be processed by other speech technology, including cloud-based technology. The technology can recognize multiple wake words that activate different digital assistants.
Sensory’s embedded speech recognition can recognize more commands on the device beyond wake words without using the cloud. It can also spot phrases within an utterance given a list of target phrases.
Sensory says its technologies have shipped in over three billion products from hundreds of consumer electronics manufacturers including ATT, Hasbro, Huawei, Google, Amazon, Samsung, LG, Mattel, Motorola, Plantronics, GoPro, Sony, Tencent, Garmin, LG, Microsoft, Lenovo, and more.