Speech "siglet" detection for business microscope

Jun Nishimura, Nobuo Sato, Tadahiro Kuroda

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

12 Citations (Scopus)

Abstract

"Business Microscope" is a tool which provides knowledge workers with a bird-eye view of their daily communication. To meet the problem of the energy consumption of sensor nodes and privacy concerns for wearers and non-wearers, "siglet" sensing is proposed. Siglet sensing is a way to capture very short and noise-like signals by sensors operating on a low duty ratio. To extract the useful information on workers' communication, speech siglet detection is studied. The LBG trained speech and workplace nonspeech models with Mel Frequency Cepstrum Coefficients (MFCCs) as feature vectors are utilized. A hierarchical pruning technique is studied to reduce the calculation cost of the matching process to nearly 25% and refine the classification accuracy. Our approach achieved average speech and nonspeech classification accuracy of 99.96% on 0. 1s long test siglets.

Original languageEnglish
Title of host publication6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008
Pages147-152
Number of pages6
DOIs
Publication statusPublished - 2008
Event6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008 - Hong Kong, Hong Kong
Duration: 2008 Mar 172008 Mar 21

Other

Other6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008
CountryHong Kong
CityHong Kong
Period08/3/1708/3/21

Fingerprint

Microscopes
Speech communication
Sensor nodes
worker
Industry
Energy utilization
communication
energy consumption
privacy
Communication
Sensors
workplace
Costs
costs

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Communication

Cite this

Nishimura, J., Sato, N., & Kuroda, T. (2008). Speech "siglet" detection for business microscope. In 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008 (pp. 147-152). [4517388] https://doi.org/10.1109/PERCOM.2008.83

Speech "siglet" detection for business microscope. / Nishimura, Jun; Sato, Nobuo; Kuroda, Tadahiro.

6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008. 2008. p. 147-152 4517388.

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

Nishimura, J, Sato, N & Kuroda, T 2008, Speech "siglet" detection for business microscope. in 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008., 4517388, pp. 147-152, 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008, Hong Kong, Hong Kong, 08/3/17. https://doi.org/10.1109/PERCOM.2008.83
Nishimura J, Sato N, Kuroda T. Speech "siglet" detection for business microscope. In 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008. 2008. p. 147-152. 4517388 https://doi.org/10.1109/PERCOM.2008.83
Nishimura, Jun ; Sato, Nobuo ; Kuroda, Tadahiro. / Speech "siglet" detection for business microscope. 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008. 2008. pp. 147-152
@inproceedings{6f2e37d5a5254e28b3fd07242e4a4cf0,
title = "Speech {"}siglet{"} detection for business microscope",
abstract = "{"}Business Microscope{"} is a tool which provides knowledge workers with a bird-eye view of their daily communication. To meet the problem of the energy consumption of sensor nodes and privacy concerns for wearers and non-wearers, {"}siglet{"} sensing is proposed. Siglet sensing is a way to capture very short and noise-like signals by sensors operating on a low duty ratio. To extract the useful information on workers' communication, speech siglet detection is studied. The LBG trained speech and workplace nonspeech models with Mel Frequency Cepstrum Coefficients (MFCCs) as feature vectors are utilized. A hierarchical pruning technique is studied to reduce the calculation cost of the matching process to nearly 25{\%} and refine the classification accuracy. Our approach achieved average speech and nonspeech classification accuracy of 99.96{\%} on 0. 1s long test siglets.",
author = "Jun Nishimura and Nobuo Sato and Tadahiro Kuroda",
year = "2008",
doi = "10.1109/PERCOM.2008.83",
language = "English",
isbn = "076953113X",
pages = "147--152",
booktitle = "6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008",

}

TY - GEN

T1 - Speech "siglet" detection for business microscope

AU - Nishimura, Jun

AU - Sato, Nobuo

AU - Kuroda, Tadahiro

PY - 2008

Y1 - 2008

N2 - "Business Microscope" is a tool which provides knowledge workers with a bird-eye view of their daily communication. To meet the problem of the energy consumption of sensor nodes and privacy concerns for wearers and non-wearers, "siglet" sensing is proposed. Siglet sensing is a way to capture very short and noise-like signals by sensors operating on a low duty ratio. To extract the useful information on workers' communication, speech siglet detection is studied. The LBG trained speech and workplace nonspeech models with Mel Frequency Cepstrum Coefficients (MFCCs) as feature vectors are utilized. A hierarchical pruning technique is studied to reduce the calculation cost of the matching process to nearly 25% and refine the classification accuracy. Our approach achieved average speech and nonspeech classification accuracy of 99.96% on 0. 1s long test siglets.

AB - "Business Microscope" is a tool which provides knowledge workers with a bird-eye view of their daily communication. To meet the problem of the energy consumption of sensor nodes and privacy concerns for wearers and non-wearers, "siglet" sensing is proposed. Siglet sensing is a way to capture very short and noise-like signals by sensors operating on a low duty ratio. To extract the useful information on workers' communication, speech siglet detection is studied. The LBG trained speech and workplace nonspeech models with Mel Frequency Cepstrum Coefficients (MFCCs) as feature vectors are utilized. A hierarchical pruning technique is studied to reduce the calculation cost of the matching process to nearly 25% and refine the classification accuracy. Our approach achieved average speech and nonspeech classification accuracy of 99.96% on 0. 1s long test siglets.

UR - http://www.scopus.com/inward/record.url?scp=49149097885&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=49149097885&partnerID=8YFLogxK

U2 - 10.1109/PERCOM.2008.83

DO - 10.1109/PERCOM.2008.83

M3 - Conference contribution

AN - SCOPUS:49149097885

SN - 076953113X

SN - 9780769531137

SP - 147

EP - 152

BT - 6th Annual IEEE International Conference on Pervasive Computing and Communications, PerCom 2008

ER -