Sixth European Conference on Speech Communication and Technology
(EUROSPEECH'99)

Budapest, Hungary
September 5-9, 1999

Topic Spotting and its Description of Summary from Spontaneous Speech

Masayuki Nakazawa, Jianxin Zhang, Ryuichi Oka

Real World Computing Partnership, Tsukuba, Ibaraki, Japan

This paper proposes two new methods of how to carry out automatic topic spotting in continuous speech and how to describe its summary in the form of slot expression. The topic spotting works independently of both speakers and topics. The method of topic spotting is based on a dialogue model to segment a topic using only a surface feature of continuous speech. We have evaluated our methods using the speech of the broadcasting news and the spoken dialogue speech. The result was that detection quality was over 80\% on the broadcasting news. This paper shows some experimental results for Japanese spontaneous speech.


Full Paper (PDF)

Bibliographic reference.  Nakazawa, Masayuki / Zhang, Jianxin / Oka, Ryuichi (1999): "Topic spotting and its description of summary from spontaneous speech", In EUROSPEECH'99, 2447-2450.