Skip to main content

Advertisement

Table 2 WERs[%] for SPINE task with different front ends/speaker normalization algorithms.

From: Towards an Intelligent Acoustic Front End for Automatic Speech Recognition: Built-in Speaker Normalization

System/WER Female Male Overall
MFCC 41.81
PMVDR 41.36
VTLN 38.28
BISN 38.25
BISN w/BTS 38.25
BISN w/MS-BTS 38.26