基于音频的电视新闻节目的主题检索和聚类

王磊; 杜利民; 王劲林

doi:10.3724/SP.J.1146.2006.00272

基于音频的电视新闻节目的主题检索和聚类

doi: 10.3724/SP.J.1146.2006.00272

计量
- 文章访问数: 3312
- HTML全文浏览量: 78
- PDF下载量: 656
- 被引次数: 0
出版历程
- 收稿日期: 2006-03-10
- 修回日期: 2006-05-30
- 刊出日期: 2007-10-19

Audio-based Topic Retrieval and Clustering of TV Broadcasting News

摘要

摘要: 随着流媒体应用的蓬勃兴起，基于媒体内容的检索和管理逐渐成为当前的学术研究热点。新闻节目作为电视节目的一种常见形式，对其主题进行自动提取检索具有重要的实际意义。该文从电视新闻节目的音频入手，综合应用了播音室语音/非播音室语音分类、说话人转换点检测以及按说话人聚类等多种技术，实现了对电视新闻节目的主题的检索和聚类。实验表明，该文中的方法能够找到新闻节目中96%以上的播音室段落，并对其进行准确归类，显示了这种方法的可行性和潜在价值。
- 新闻主题检索;音频分类;说话人检测;说话人聚类;贝叶斯信息准则
Abstract: With boosting of stream media applications, content-based media information retrieval becomes hot topic of current academic research. Since news program is familiar and popular, topic retrieval of news program has important practical significance. Based on audio processing, this paper integrates studio / non-studio classification, speaker change detection and speaker clustering, and realizes automatic news topic retrieval and clustering according to anchorman. The experiment indicates that above 96% studio segments of news programs can be found out and clustered, and proves feasibility and potential of the method.

HTML全文

参考文献(1)

Gauvain J L and Adda G. Transcribing broadcast news: the LIMSI Nov96 Hub4 System[J].Proc. ARPA Speech Recognition Workshop, Chantilly, Virginia, 1997: 56-63.[2]Cook G and Robinson T. Transcribing broadcast news with the 1997 ABBOT system. Proc. IEEE International Conference on Acoustic, Speech and Signal Processing, Seattle, 1998: 917-920.Chen S Shaobing and Gopalakrishnan P S. Speaker, environment and channel change detection and clustering via the Bayesian information criterion. Proc. DARPA Broadcast News Transcription and Understanding Workshop, Landsdowne, VA, 1998: 127-132.[3]Delacourt P, Kryze D, and Wellekens C J. Speaker-based segmentation for audio data indexing. Proc. ESCA-ETRW workshop on Accessing Information in Spoken Audio, Cambridge, UK, 1999: 1195-1198.[4]Solomonoff A, Mielke A, Schmidt M, and Gish H. Clustering speakers by their voices. Proc. IEEE International Conference on Acoustic, Speech and Signal Processing, Seattle, 1998: 757-760.Reynolds D A.[J].Singer E, Carlson B A, Orsquo;Leary G C, Mclaughlin J J, and Zissman M A. Blind clustering of speech utterances based on speaker and language characteristics. Proc. the International Conference on Speech and Language Processing, Sydney.1998,:-[5]杨行峻，迟惠生. 语音信号数字处理. 第一版. 北京：电子工业出版社，1995，第4章.[6]Ajmera J and McCowan I. Robust speaker change detection. IEEE Signal Processing Letters.2004, 11(8):649-651.[7]Couvreur L and Boite J M. Speaker tracking in broadcast audio material in the framework of the THISL project. Proc. ESCA-ETRW Workshop on Accessing Information in Spoken Audio, Cambridge, UK, 1999: 84-89.[8]Iurgel U and Meermeier R. New approaches to audic-visual segmentation of TV news for automatic topic retrieval. Proc. IEEE International Conference on Acoustic, Speech and Signal Processing, Salt Lake City, Utah, 2001: 1397-1400.

施引文献

资源附件(0)

访问统计