Machine Listening in Multisource Environments (CHiME) 2011

Florence, Italy
September 1, 2011

A Two-Channel Acoustic Front-End for Robust Automatic Speech Recognition in Noisy and Reverberant Environments

Roland Maas, Andreas Schwarz, Yuanhang Zheng, Klaus Reindl, Stefan Meier, Armin Sehr, Walter Kellermann

Multimedia Communications and Signal Processing, University of Erlangen-Nuremberg, Germany

An acoustic front-end for robust automatic speech recognition in noisy and reverberant environments is proposed in this contribution. It comprises a blind source separation-based signal extraction scheme and only requires two microphone signals. The proposed front-end and its integration into the recognition system is analyzed and evaluated in noisy living room-like environments according to the PASCAL CHiME challenge. The results show that the introduced system significantly improves the recognition performance compared to the challenge baseline.

Index Terms. PASCAL CHiME challenge, robust automatic speech recognition, blind source extraction, speech enhancement

Full Paper

Bibliographic reference.  Maas, Roland / Schwarz, Andreas / Zheng, Yuanhang / Reindl, Klaus / Meier, Stefan / Sehr, Armin / Kellermann, Walter (2011): "A two-channel acoustic front-end for robust automatic speech recognition in noisy and reverberant environments", In CHiME-2011, 41-46.