Table 3 Global accuracies of the HMM-GMM-, CNN-, and EMDH-CNN-based dysarthric speech recognition systems
| Â | Systems | |||
|---|---|---|---|---|
All speakers | HMM-GMM | CNN | EMDH-CNN (70/30) | EMDH-CNN 10-fold |
Accuracy (%) | 44.13% | 54.91% | 57.54% | 64.86% |
| Â | Systems | |||
|---|---|---|---|---|
All speakers | HMM-GMM | CNN | EMDH-CNN (70/30) | EMDH-CNN 10-fold |
Accuracy (%) | 44.13% | 54.91% | 57.54% | 64.86% |