SpeeChin, the device to give silent voice commands

Although not everyone likes them, virtual assistants are used in countless tasks, whether through mobile phones, cars, smart speakers, home devices, etc.

Siri, Alexa, Google Assistant, etc., receive voice commands. But this has an obvious limitation: you can’t use it in a place where you have to be quiet. Or on the contrary, in a place with a lot of noise. Or by a mute person, or who has voice problems.

Ruidong Zhang, a computer science student at Cornell University, in the United States, with the help of assistant professor Cheng Zhang, has developed SpeeChin, a device to give silent voice commands. You can see it in this video:

SpeeChin is used to give voice commands to a virtual assistant,without making sounds.

It is an infrared camera placed on a collar, and oriented to point at the chin. Although its function is to read lips, it is placed on the chin so as not to focus on the face, preserving privacy.

It’s more private, but that makes command recognition much more complicated.

SpeeChin employs artificial intelligence that has been trained to recognize voice commands, through chin movements.

What to look for before buying an Amazon Echo smart speaker with Alexa

When we pronounce a word, our chin and Adam’s apple move differently depending on the letters that make up the word.

SpeeChin’s artificial intelligence has been trained by 10 students who spoke English, and another ten students who spoke Chinese, speaking the most common commands from Alexa and the Google Assistant, silently.

After training, this AI is capable of recognizing 54 commands in English and 44 commands in Chinese, with a success rate of 90.5 and 91.6%, respectively. Just by focusing the camera on the chin, without seeing the mouth.

It is an interesting technology, because it is cheap, simple to manufacture, and respects privacy. It can be very useful in specific jobs where there is a lot of noise, or on the contrary, where total silence is required.