Researchers from Tsinghua University and Shanghai Jiao Tong University in China have invented a smart artificial throat (AT) that is sensitive to speech or movements related to articulation, which is expected to assist patients who have undergone laryngectomy.
The smart artificial throat can recognize chords and mechanical movements. (Illustration: Createdigital).
In this study, the experts placed a laser-cut graphene material on a layer of polyvinyl alcohol film.
Measuring approximately 0.6 inches by 1.2 inches—twice the size of a thumb—the artificial throat can be attached to the patient’s neck.
The research team used water to adhere the film to the throat of the patient and then connected it to an electrode and a microcomputer, power amplifier, and motion decoder.
The study, published in the journal Nature Machine Intelligence, shows that the AT can recognize chords and mechanical movements.
This capability allows the AT to capture signals with low fundamental frequencies and has noise immunity.
The research indicates that the AT can detect fundamental speech elements such as phonemes, tones, and words with an average accuracy of up to 99%.
Through a comprehensive artificial intelligence (AI) model, the AT can recognize everyday words from laryngectomy patients with over 90% accuracy.
The content of the spoken words will be synthesized into complete speech and emitted from the AT, helping patients gradually regain their ability to articulate after surgery.
The research team noted that there are still many aspects to optimize, such as sound quality, volume, and voice diversity, and they hope that in the future, this technology can help voice-loss patients regain their ability to speak.
