CNN-based Multichannel End-to-End Speech Recognition for Everyday Home Environments

    •  Yalta, N., Watanabe, S., Hori, T., Nakadai, K., Ogata, T., "CNN-based Multichannel End-to-End Speech Recognition for Everyday Home Environments", European Signal Processing Conference (EUSIPCO), DOI: 10.23919/EUSIPCO.2019.8902524, September 2019, pp. 1-5.
      BibTeX TR2019-094 PDF
      • @inproceedings{Yalta2019sep,
      • author = {Yalta, Nelson and Watanabe, Shinji and Hori, Takaaki and Nakadai, Kazuhiro and Ogata, Tetsuya},
      • title = {CNN-based Multichannel End-to-End Speech Recognition for Everyday Home Environments},
      • booktitle = {European Signal Processing Conference (EUSIPCO)},
      • year = 2019,
      • pages = {1--5},
      • month = sep,
      • doi = {10.23919/EUSIPCO.2019.8902524},
      • url = {}
      • }
  • MERL Contact:
  • Research Areas:

    Artificial Intelligence, Machine Learning, Speech & Audio

Casual conversations involving multiple speakers and noises from surrounding devices are common in everyday environments, which degrades the performances of automatic speech recognition systems. These challenging characteristics of environments are the target of the CHiME-5 challenge. By employing a convolutional neural network (CNN)-based multichannel end-to-end speech recognition system, this study attempts to overcome the presents difficulties in everyday environments. The system comprises of an attention-based encoder–decoder neural network that directly generates a text as an output from a sound input. The multichannel CNN encoder, which uses residual connections and batch renormalization, is trained with augmented data, including white noise injection. The experimental results show that the word error rate is reduced by 8.5% and 0.6% absolute from a single channel endto-end and the best baseline (LF-MMI TDNN) on the CHiME-5 corpus, respectively.