Voice rehabilitation and diseases prediction is required today because neural degeneration or neurological injury alters the motor component of the speech system in the phonation area of the brain. A novel approach to voice rehabilitation consists in predicting the phonetic control by the EMG. In a previous work we demonstrated that the voice-production apparatus (tongue muscle) generates a specific EMG signal that identify the phoneme emitted. The inference paradigm is EFuNN (Evolving Fuzzy Neural Network) trained by the sampled EMG (Electro Myo Gram) signal at phonation-. time. A phoneme-to-speech non-invasive epidermal patch is to be designed with energy harvestimg and MEMS loudspeakers. |
*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.