Skip to main content

Table 9 WRE approach

From: Compact Acoustic Models for Embedded Speech Recognition

  Without adaptation With adaptation
  MLE FDW FD MLE FDW FD
Very compact model (6 k.) 6,05% 8,54% 5,99% 5,48% 8,67% 5,36%
Compact model (11 k.) 5,15% 7,50% 5,15% 4,67% 7,28% 4,63%