Research Article

Estimation of Phoneme-Specific HMM Topologies for the Automatic Recognition of Dysarthric Speech

Figure 4

Ergodic structure of a HMM for modelling of words.
297860.fig.004