An Unsupervised Hybrid Approach for Online Detection of Sound Scene Changes in Broadcast Content

Alexandra Craciun, Gökhan Sevkin, Tom Bäckström

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Abstract

In this paper we describe an online system for broadcast content, which can detect sound scene changes with high accuracy. The system is unsupervised and does not require prior information on the segment classes. A scene change probability score is computed for each frame of the signal using a hybrid approach combining a model-based (Gaussian Mixture Model) with a distance-based (Hotelling's T2-Statistic) segmentation method. The mixture model parameters are adapted online using the previous frames of the signal. Experiments on real recordings show that we can achieve more than 85% correct segment change detection with only 16% false detections.
Original languageEnglish
Title of host publicationProceedings of the AES Conference on Semantic Audio
EditorsChristian Dittmar, Jakob Abeßer , Meinard Müller
PublisherAudio Engineering Society
Pages53-60
ISBN (Electronic)978-1-942220-15-2
Publication statusPublished - Jun 2017
MoE publication typeA4 Article in a conference publication
EventAES International Conference on Semantic Audio - Erlangen, Germany
Duration: 22 Jun 201724 Jun 2017
http://www.aes.org/conferences/2017/semantic/

Conference

ConferenceAES International Conference on Semantic Audio
CountryGermany
CityErlangen
Period22/06/201724/06/2017
Internet address

Fingerprint

Dive into the research topics of 'An Unsupervised Hybrid Approach for Online Detection of Sound Scene Changes in Broadcast Content'. Together they form a unique fingerprint.

Cite this