Title :
Exploring multi-channel features for denoising-autoencoder-based speech enhancement
Author :
Araki, Shoko ; Hayashi, Tomoki ; Delcroix, Marc ; Fujimoto, Masakiyo ; Takeda, Kazuya ; Nakatani, Tomohiro
Author_Institution :
NTT Commun. Sci. Labs., NTT Corp., Kyoto, Japan
Abstract :
This paper investigates a multi-channel denoising autoencoder (DAE)-based speech enhancement approach. In recent years, deep neural network (DNN)-based monaural speech enhancement and robust automatic speech recognition (ASR) approaches have attracted much attention due to their high performance. Although multi-channel speech enhancement usually outperforms single channel approaches, there has been little research on the use of multi-channel processing in the context of DAE. In this paper, we explore the use of several multi-channel features as DAE input to confirm whether multi-channel information can improve performance. Experimental results show that certain multi-channel features outperform both a monaural DAE and a conventional time-frequency-mask-based speech enhancement method.
Keywords :
interference suppression; speech coding; speech enhancement; ASR; DNN-based monaural speech enhancement; deep neural network-based monaural speech enhancement; multi-channel DAE-based speech enhancement approach; multi-channel denoising autoencoder-based speech enhancement approach; multi-channel information; multi-channel processing; multi-channel speech enhancement; robust automatic speech recognition approaches; Artificial neural networks; Filter banks; Noise reduction; Testing; Training; Deep learning; PASCAL ‘CHiME’ challenge; denoising autoencoder; multi-channel noise suppression;
Conference_Titel :
Acoustics, Speech and Signal Processing (ICASSP), 2015 IEEE International Conference on
Conference_Location :
South Brisbane, QLD
DOI :
10.1109/ICASSP.2015.7177943