Locally Normalized Filter Banks Applied to Deep Neural-Network-Based Robust Speech Recognition

This letter describes modifications to locally normalized filter banks (LNFB), which substantially improve their performance on the Aurora-4 robust speech recognition task using a Deep Neural Network-Hidden Markov Model (DNN-HMM)-based speech recognition system. The modified coefficients, referred t...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	IEEE signal processing letters 2017-04, Vol.24 (4), p.377-381
Hauptverfasser:	Fredes, Josue, Novoa, Jose, King, Simon, Stern, Richard M., Becerra Yoma, Nestor
Format:	Artikel
Sprache:	eng
Schlagworte:	Aurora-4 Automatic speech recognition (ASR) channel mismatch deep neural network (DNN) Filter banks Histograms locally normalized filter bank (LNFB) Manganese Mel frequency cepstral coefficient Robustness Testing Training
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!