DeepRESTORE: Deep learning approaches for speech restoration from face movement biosignals

Periodo: desde 2023 hasta 2026

Entidad financiadora: Ministerio de Ciencia e Innovación

Importe total: 186875€

Investigadores Principales: HERNAEZ RIOJA, INMACULADA SARATXAGA COUCEIRO / IBON

Descripción:

The GENERAL OBJECTIVE of DeepRESTORE is to overcome the limitations of both traditional voice rehabilitation methods and previous SSI studies by investigating SSI-based communication systems for restoring communication to individuals who have been deprived of the ability to speak. In particular, in the subproject that will be developed at UPV/EHU total-laryngectomy patients, whose upper vocal tract (i.e. mouth, tongue, lips, etc.) is intact but cannot speak normally because their entire larynx (an organ essential for sound production) has been surgically removed to treat laryngeal cancer will be considered. These persons still retain the control over their speech articulators and, hence, biosignals reflecting the movements of the articulators can be easily captured using EMG and or video images of their faces.

Previous post Aholab in the media: La mecánica del caracol Next post ILENIA: Impulso de las lenguas en la inteligencia artificial