Abstract
This paper analyses the inter-experimental similarities in the muscle activation during vowel sound production by an individual. Surface electromyography has been used as an indicator of muscle activity and independent component analysis has been used to separate the electrical activity from different muscles. The results indicate that there is a 'reasonable' relationship between muscle activities of the corresponding muscles when the experiments are repeated. The results demonstrate that when people speak, they use a similar set of muscles when they repeat the same sound. The results also indicate that there is a variation when the same sound is spoken at different speeds of utterances. This can be attributable to the lack of audio feedback when the same sound is uttered.
Original language | English |
---|---|
Pages (from-to) | 215-223 |
Number of pages | 9 |
Journal | Computer Methods in Biomechanics and Biomedical Engineering |
Volume | 13 |
Issue number | 2 |
DOIs | |
Publication status | Published - 2010 |
Keywords
- blind source separation
- electromyography
- face
- independent component analysis
- muscles
- neural networks (computer science)
- speech