Multi-view features in a DNN-CRF model for improved sentence unit detection on English broadcast news

Guangpu Huang, Chenglin Xu, Xiong Xiao, Lei Xie, Eng Siong Chng, Haizhou Li

科研成果: 书/报告/会议事项章节会议稿件同行评审

5 引用 (Scopus)

摘要

This paper presents a deep neural network-conditional random field (DNN-CRF) system with multi-view features for sentence unit detection on English broadcast news. We proposed a set of multi-view features extracted from the acoustic, articulatory, and linguistic domains, and used them together in the DNN-CRF model to predict the sentence boundaries. We tested the accuracy of the multi-view features on the standard NIST RT-04 English broadcast news speech data. Experiments show that the best system outperforms the state-of-the-art sentence unit detection system significantly by 13.2% absolute NIST sentence error rate reduction using the reference transcription. However, the performance gain is limited on the recognized transcription partly due to the high word error rate.

源语言英语
主期刊名2014 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA 2014
出版商Institute of Electrical and Electronics Engineers Inc.
ISBN(电子版)9786163618238
DOI
出版状态已出版 - 12 2月 2014
活动2014 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA 2014 - Chiang Mai, 泰国
期限: 9 12月 201412 12月 2014

出版系列

姓名2014 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA 2014

会议

会议2014 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA 2014
国家/地区泰国
Chiang Mai
时期9/12/1412/12/14

指纹

探究 'Multi-view features in a DNN-CRF model for improved sentence unit detection on English broadcast news' 的科研主题。它们共同构成独一无二的指纹。

引用此