Restoring speech following total removal of the larynx by a learned transformation from sensor data to acoustics

Gilbert, James M.; Gonzalez, Jose A.; Cheah, Lam A.; Ell, Stephen R.; Green, Phil; Moore, Roger K.; Holdsworth, Ed

Abstract

Total removal of the larynx may be required to treat laryngeal cancer: speech is lost. This article shows that it may be possible to restore speech by sensing movement of the remaining speech articulators and use machine learning algorithms to derive a transformation to convert this sensor data into an acoustic signal. The resulting “silent speech,” which may be delivered in real time, is intelligible and sounds natural. The identity of the speaker is recognisable. The sensing technique involves attaching small, unobtrusive magnets to the lips and tongue and monitoring changes in the magnetic field induced by their movement.

Citation

Gilbert, J. M., Gonzalez, J. A., Cheah, L. A., Ell, S. R., Green, P., Moore, R. K., & Holdsworth, E. (2017). Restoring speech following total removal of the larynx by a learned transformation from sensor data to acoustics. The Journal of the Acoustical Society of America, 141(3), EL307-EL313. https://doi.org/10.1121/1.4978364

• Journal Article Type: Article
• Acceptance Date: Feb 21, 2017
• Online Publication Date: Mar 21, 2017
• Publication Date: 2017-03
• Deposit Date: Mar 28, 2017
• Publicly Available Date: Dec 5, 2018
• Journal: Journal of the Acoustical Society of America
• Print ISSN: 0001-4966
• Publisher: Acoustical Society of America
• Peer Reviewed: Peer Reviewed
• Volume: 141
• Issue: 3
• Pages: EL307-EL313
• DOI: https://doi.org/10.1121/1.4978364
• Keywords: Speech restoration; Permanent magnetic articulography; Machine learning
• Public URL: https://hull-repository.worktribe.com/output/450047
• Publisher URL: http://asa.scitation.org/doi/10.1121/1.4978364
• Related Public URLs: http://eprints.whiterose.ac.uk/114030/
• Additional Information: received: 2016-10-17; revised: 2017-01-22; accepted: 2017-02-21; published: 2017-03-21
• Contract Date: Mar 28, 2017