Cornell Researchers Develop AI-Based Eyeglasses To Decipher Silent Speech

Artificial intelligence has witnessed a notable surge in prominence within the healthcare sector, particularly in recent times. And researchers at Cornell University have come up with a new breakthrough that they believe could revolutionize the field once again. 

Recently, the researchers unveiled EchoSpeech – a wearable, low-powered device that operates using AI and acoustic sensing to recognize unvocalized commands based on nothing but mouth and lip movements.

EchoSpeech is capable of comprehending and identifying commands with just a few minutes of user training data, making it highly efficient. Moreover, it can seamlessly integrate into smartphones, further enhancing its versatility.

In a paper, Ruidong Zhang, the lead researcher, explained that the product could be a perfect voice synthesizer that will be used by people who can’t vocalize. In its present form, EchoSpeech can help people to communicate via smartphones in places where vocalized speech might be impossible or inconvenient – libraries, noisy places, etc. 

The researchers have added a pair of speakers and microphones to the glasses, making them a wearable AI-powered sonar device. Its job will be to receive and send sound waves across the face, sensing movements of the mouth. And then, a deep learning algorithm will analyze the echo profiles in real time, with the researchers claiming that it has a 95% accuracy rate.

These wearables could easily be a revolution. So far, technology in silent-speech recognition has been mostly limited to a few predetermined commands – while users have been required to wear or face a camera. Besides that, Zhang also pointed out that there are privacy concerns with wearable cameras. 

Leveraging acoustic sensing technology, EchoSpeech eliminates the requirement for wearable video cameras. And since the audio data is significantly smaller, it won’t need much bandwidth to process.