‘Smart choker’ recognizes words with 95% accuracy

Cambridge University researchers have developed a smart choker that uses artificial intelligence (AI) to help people with speech impairment to communicate.

The device can detect the tiniest of movements in the throat, and tests show that it can decode speech with 95.2% accuracy, according to Cambridge University.

“I was surprised at just how sensitive the device is,” said Dr. Luigi Occhipinti from the Cambridge Graphene Centre, who led the research.

“We couldn’t capture all the signals and complexity of human speech before, but now that we can, it unlocks a whole new set of potential applications,” Occhipinti said.

The device could be useful for people who have temporary or permanent speech impairments as a result of laryngeal surgery or conditions like Parkinson’s, stroke, or cerebral palsy.

“Current solutions for people with speech impairments often fail to capture words and require a lot of training,” Occhipinti said, adding, “They are also rigid, bulky and sometimes require invasive surgery to the throat.”

The choker analyzes non-vocal signals to decode speech in silent conditions – a type of technology known as a silent speech interface. This means the user only needs to mouth the words in order for them to be captured.

The choker has sensors that are more than four times higher in sensitivity than similar technology so far – all incorporated in a soft, stretchable fabric that makes it comfortable to wear.

“These sensors can detect tiny vibrations, such as those formed in the throat when whispering or even silently mouthing words, which makes them ideal for speech detection,” Occhipinti continued.

Captured speech signals can then be transferred to a computer or speaker to facilitate conversation. The device uses a combination of flexible electronics and machine learning techniques to recognize differences in pronunciation, accent, and vocabulary between users.

Researchers said the choker will have to undergo “extensive” testing and clinical trials before it can be approved for use in patients with speech impairments.

Beyond the original purpose, it could also be used in other health monitoring applications or to improve communication in noisy or secure environments, they said.

The results of the study were reported in the journal npj Flexible Electronics.