Skip to main content

Table 7 Comparison of phone-class features (WER (%))

From: Reverberant speech recognition combining deep neural networks and deep autoencoders augmented with a phone-class feature

Recognition Training

P C soft

\(PC^{oracle}_{\textit {hard}}\)

P C soft

23.29

24.10

P C hard

23.27

24.34

\(PC^{decode}_{\textit {hard}}\)

22.85

23.29

(cf.) \(PC^{oracle}_{\textit {hard}}\)

13.74

14.25