CNN-based multichannel end-to-end speech recognition for everyday home environments

Nelson Yalta, Shinji Watanabe, Takaaki Hori, Kazuhiro Nakadai, Tetsuya Ogata

研究成果: Conference contribution

3 被引用数 (Scopus)

抄録

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 end-to-end and the best baseline (LF-MMI TDNN) on the CHiME-5 corpus, respectively.

本文言語English
ホスト出版物のタイトルEUSIPCO 2019 - 27th European Signal Processing Conference
出版社European Signal Processing Conference, EUSIPCO
ISBN(電子版)9789082797039
DOI
出版ステータスPublished - 2019 9
イベント27th European Signal Processing Conference, EUSIPCO 2019 - A Coruna, Spain
継続期間: 2019 9 22019 9 6

出版物シリーズ

名前European Signal Processing Conference
2019-September
ISSN(印刷版)2219-5491

Conference

Conference27th European Signal Processing Conference, EUSIPCO 2019
国/地域Spain
CityA Coruna
Period19/9/219/9/6

ASJC Scopus subject areas

  • 信号処理
  • 電子工学および電気工学

フィンガープリント

「CNN-based multichannel end-to-end speech recognition for everyday home environments」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル