Sound-event partitioning and feature normalization for robust sound-event detection

Baiying Lei, Man Wai Mak

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

5 Citations (Scopus)

Abstract

The ubiquitous of smartphones has opened up the possibility of mobile acoustic surveillance. However, the continuous operation of surveillance systems calls for efficient algorithms to conserve battery consumption. This paper proposes a power-efficient sound-event detector that exploits the redundancy in the sound frames. This is achieved by a soundevent partitioning (SEP) scheme where the acoustic vectors within a sound event are partitioned into a number of chunks, and the means and standard deviations of the acoustic features in the chucks are concatenated for classification by a support vector machine (SVM). Regularized PCA-whitening and L2 normalization are applied to the acoustic vectors to make them more amenable for the SVM. Experimental results based on 1000 sound events show that the proposed scheme is effective even if there are severe mismatches between the training and test conditions.
Original languageEnglish
Title of host publication2014 19th International Conference on Digital Signal Processing, DSP 2014
PublisherIEEE
Pages389-394
Number of pages6
Volume2014-January
ISBN (Electronic)9781479946129
DOIs
Publication statusPublished - 1 Jan 2014
Event2014 19th International Conference on Digital Signal Processing, DSP 2014 - Hong Kong, Hong Kong
Duration: 20 Aug 201423 Aug 2014

Conference

Conference2014 19th International Conference on Digital Signal Processing, DSP 2014
CountryHong Kong
CityHong Kong
Period20/08/1423/08/14

Keywords

  • Feature normalization
  • PCA whitening and regularization
  • Scream sound detection
  • Sound event partitioning

ASJC Scopus subject areas

  • Signal Processing

Cite this