Cornell University is proud to present its brand new concept: glasses capable of reading and detecting the movements of the mouth.
And the project is already at a very advanced stage of its design.
Sonar glasses that follow the movements of your face
It is mainly to Ruidong Zhang, a researcher at Cornell University, that we owe this new concept of glasses. On paper, the principle is simple: to offer you a discreet way to control your smartphone without having to touch the screen. This pair of glasses has tiny microphones and speakers to detect the words you speak silently. Here is what the researchers from the University explain:
“It’s not telepathy: it’s the seemingly ordinary, out-of-the-box glasses he’s wearing, called EchoSpeech – a silent voice-recognition interface that uses acoustic detection and artificial intelligence to continuously recognize up to 31 non-voice commands, based on lip and mouth movements. »
According to the researchers, the system would only take a few minutes to be able to recognize the user’s speech patterns. The latter can then pronounce the password of his phone in order to unlock it or even launch the next song from his playlist. To work, the technology uses a learning algorithm that analyzes echo profiles with a claimed accuracy of around 95%.
A technological advance, but not only
Cornell University considers this concept to be a step forward in privacy. The data is in fact processed locally on the smartphone and is therefore not uploaded to the cloud. In addition, EchoSpeech eliminates the need to carry a camera, thus allowing you to remain more discreet.
According to Zhang, the potential of these glasses goes far beyond the issue of privacy. The researcher considers that the technology incorporated therein is also aimed at people with speech impairments: “ For people who can’t vocalize sound, this silent speech technology could be a great input for a speech synthesizer. This could give voice back to patients “.
The battery offers a life of 10 hours for acoustic detection. The Smart Computing Interfaces for Future Interactions (SciFi) Lab is currently exploring the possibility of commercializing the technology behind EchoSpeech with the help of a funding program from Cornell. Apps to track facial, eye and upper body movements are also being explored.
Source : Engadget
7