Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing

Table 4 ASR performance in WERs of MFCCs and their combination with FDLPs modeled by HMM/GMM-, hybrid (HMM/DNN)-, and BN-HMM/GMM-based systems on MP-FR evaluation data

From: Exploiting foreign resources for DNN-based ASR

System

HMM/GMM (%)

HMM/DNN (%)

BN-HMM/GMM (%)

MFCC

17.1

14.9

14.5

MFCC+FDLP

16.5

14.4

13.9

  1. The acoustic models are trained on MP-FR (19h) data
Back to article page