Haar-like filtering based speech detection using integral signal for sensornet

Jun Nishimura, Tadahiro Kuroda

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

5 Citations (Scopus)

Abstract

Speech detection using haar-like filtering is proposed as a new and very low calculation cost method for sensornet applications. The simple haar-like filters having variable filter width and shift width are trained to learn appropriate filter parameters from the training samples to detect speech. To further decrease the calculation cost, the use of intermediate signal representation called "integral signal" is proposed. Our method yielded speech/nonspeech classification accuracy of 97.44% for the input length of 0.1s. Compared with high performance feature extraction method MFCC (Mel-Frequency Cepstrum Coefficient), the proposed haar-like filtering can be approximately 93.71% efficient in terms of the total amount of add and multiply calculations while capable of achieving the error rate of only 2.56% relative to MFCC.

Original languageEnglish
Title of host publicationProceedings of the 3rd International Conference on Sensing Technology, ICST 2008
Pages52-56
Number of pages5
DOIs
Publication statusPublished - 2008
Externally publishedYes
Event3rd International Conference on Sensing Technology, ICST 2008 - Tainan, Taiwan, Province of China
Duration: 2008 Nov 302008 Dec 3

Publication series

NameProceedings of the 3rd International Conference on Sensing Technology, ICST 2008

Other

Other3rd International Conference on Sensing Technology, ICST 2008
Country/TerritoryTaiwan, Province of China
CityTainan
Period08/11/3008/12/3

Keywords

  • Haar-like filtering
  • Integral signal
  • Sensornet
  • Speech detection

ASJC Scopus subject areas

  • Software
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Haar-like filtering based speech detection using integral signal for sensornet'. Together they form a unique fingerprint.

Cite this