Robust speech recognition using DNN-HMM acoustic model combining noise-aware training with spectral subtraction.

Akihiro Abe, Kazumasa Yamamoto, Seiichi Nakagawa

Browse the full Interspeech paper archive.