Embedded Wake Word & Phrase Recognition Engine

Always-Listening Wake Word

A wake word is the gateway between you and your digital assistant. Voice assistants, such as Alexa and Siri, are powered by AI with word detection abilities that enable them to respond to queries and commands. Common wake words include “Hey, Google, Alexa, and Hey Siri.” Today’s wake word performance and speech recognition are made possible through the use of machine learning or AI in tandem with cloud processing. Sensory’s wake word and phrase recognition engine goes further and leverages the use of deep neural networks to provide an all embedded or on-device wake word and phrase recognition engine

Why Embedded Word Detection?

The advantage to using an embedded solution is that the voice recognition engine is trained to work without reliance on a wifi connection which results in a solution that is highly accurate under a multitude of conditions. Additionally, attempts at uttering the wake word or commands are never sent to the cloud, virtually eliminating privacy concerns. Sensory’s embedded wake word technology has been used in cases that extend well beyond mobile applications. The low power feature of TrulyHandsfree™ enables its use in battery powered devices like Bluetooth headphones, smart watches, cameras, and emergency alert devices for the aging population. The phrase spotting technology component of TrulyHandsfree™ (THF) allows users to utter commands in a natural way to control devices beyond just wake up. Queries like “what time is it”, or “how many steps have I taken” are examples of phrases that THF can process with high accuracy and zero latency. Sensory’s wake word technology may be integrated with our full suite of voice recognition applications to enable touch free food ordering at a kiosk, voice controlled microwaves, or even user identification for settings in a tv or vehicle.

How Wake Words Work

Wake words, sometimes dubbed “hotwords” are not created equal, and this is important because a wake word can be considered the key to the front door. When the wake word is uttered, the door should open every time, in other words the rate of false rejects is ideally close to zero. When the wake word is NOT said the door shouldn’t open, or the rate of false accepts should also be close to zero. Achieving this balance of low false accepts and false rejects in an embedded algorithm for a multitude of applications requires well trained and fine-tuned AI, and with over 25 years of experience, Sensory can deliver.

Highly accurate hotword detection is made complex by factors like noise, distance, and variations in accents and language. Platforms such as Amazon, Google, and a few others seem to have pretty good wake words, but if you go into your Alexa settings you can see all of the voice data that’s been collected, and a lot of it is being collected when you weren’t intentionally talking to Alexa! You can see this performance issue in the Vocalize test report. Sensory substantially outperformed Amazon in the false reject area. There’s a reason that companies like Amazon, Google, Microsoft and Samsung have turned to Sensory for our TrulyHandsfree technology.

Learn more about wake word performance in this evaluation.