HMM-based acoustic event detection with adaboost feature selection

Xi Zhou, Xiaodan Zhuang, Ming Liu, Hao Tang, Mark Hasegawa-Johnson, Thomas Huang

Research output: Contribution to journalConference article

Abstract

Because of the spectral difference between speech and acous- tic events, we propose using Kullback-Leibler distance to quantify the discriminant capability of all speech feature components in acoustic event detection. Based on these distances, we use AdaBoost to select a discriminant feature set and demonstrate that this feature set outperforms classical speech feature set such as MFCC in one-pass HMM-based acoustic event detection. We implement an HMM-based acoustic events detection system with lattice rescoring using a feature set selected by the above AdaBoost based approach.

Original languageEnglish (US)
Pages (from-to)345-353
Number of pages9
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4625 LNCS
DOIs
StatePublished - Jul 28 2008
Event2nd Annual Classifcation of Events Activities and Relationships, CLEAR 2007 and Rich Transcription, RT 2007 - Baltimore, MD, United States
Duration: May 8 2007May 11 2007

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'HMM-based acoustic event detection with adaboost feature selection'. Together they form a unique fingerprint.

  • Cite this