Audio source segmentation using spectral correlation features for automatic indexing of broadcast news

Shoichi Matsunaga, Osamu Mizuno, Katsutoshi Ohtsuki, Yoshihiko Hayashi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper proposes a new segmentation procedure to detect audio source intervals for automatic indexing of broadcast news. The procedure is composed of an audio source detection part and a part that smoothes the detected sequences. The detection part uses three new acoustic feature parameters that are based on spectral cross-correlation: spectral stability, white noise similarity, and sound spectral shape. These parameters make it possible to capture the audio sources more accurately than can be done with conventional parameters. The smoothing part has a new merging method that drops erroneous detection results of short duration. Audio source classification experiments are conducted on broadcast news segments. Performance is increased by 6.6% when the proposed parameters are used and by 3.1% when the proposed merging method is used, showing the usefulness of our approach. Experiments confirm the impact of this proposal on broadcast news indexing.

Original languageEnglish
Title of host publication2004 12th European Signal Processing Conference, EUSIPCO 2004
PublisherEuropean Signal Processing Conference, EUSIPCO
Pages2103-2106
Number of pages4
ISBN (Electronic)9783200001657
Publication statusPublished - 2015 Apr 3
Externally publishedYes
Event12th European Signal Processing Conference, EUSIPCO 2004 - Vienna, Austria
Duration: 2004 Sep 62004 Sep 10

Publication series

NameEuropean Signal Processing Conference
Volume06-10-September-2004
ISSN (Print)2219-5491

Conference

Conference12th European Signal Processing Conference, EUSIPCO 2004
CountryAustria
CityVienna
Period04/9/604/9/10

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Audio source segmentation using spectral correlation features for automatic indexing of broadcast news'. Together they form a unique fingerprint.

  • Cite this

    Matsunaga, S., Mizuno, O., Ohtsuki, K., & Hayashi, Y. (2015). Audio source segmentation using spectral correlation features for automatic indexing of broadcast news. In 2004 12th European Signal Processing Conference, EUSIPCO 2004 (pp. 2103-2106). [7079871] (European Signal Processing Conference; Vol. 06-10-September-2004). European Signal Processing Conference, EUSIPCO.