Robust speech recognition in unknown reverberant and noisy conditions

Roger Hsiao, Jeff Z. Ma, William Hartmann, Martin Karafiát, Frantisek Grézl, Lukás Burget, Igor Szöke, Jan Cernocký, Shinji Watanabe, Zhuo Chen, Sri Harish Reddy Mallidi, Hynek Hermansky, Stavros Tsakalidis, Richard M. Schwartz

Abstract

In this paper, we describe our work on the ASpIRE (Automatic Speech recognition In Reverberant Environments) challenge, which aims to assess the robustness of automatic speech recognition (ASR) systems. The main characteristic of the challenge is developing a high-performance system without access to matched training and development data. While the evaluation data are recorded with far-field microphones in noisy and reverberant rooms, the training data are telephone speech and close talking. Our approach to this challenge includes speech enhancement, neural network methods and acoustic model adaptation, We show that these techniques can successfully alleviate the performance degradation due to noisy audio and data mismatch.
BibTeX
@inproceedings{hsiao2015robust,
  author = {Roger Hsiao and Jeff Z. Ma and William Hartmann and Martin Karafiát and Frantisek Grézl and Lukás Burget and Igor Szöke and Jan Cernocký and Shinji Watanabe and Zhuo Chen and Sri Harish Reddy Mallidi and Hynek Hermansky and Stavros Tsakalidis and Richard M. Schwartz},
  title = {{Robust speech recognition in unknown reverberant and noisy conditions}},
  booktitle = {ASRU},
  year = {2015},
  doi = {10.1109/asru.2015.7404841},
}

Loading related content...

Figure