Presentation 2012-06-28
Multi-Class EMG Pattern Classfication using a Selective Desensitization Neural Network with a Half-vs-Half Method
Kazumasa HORIE, Atsuo SUEMITSU, Masahiko MORITA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The real-time classification of human movements by using surface electromyogram (EMG) signals is an important research issue in the development of a new input interface. However, the existing approaches detect irrelevant movements as the target movement, that is, false detection, if the number of the movements to be classified is large. In this research, we propose a new multi-class classification method (half-vs-half method) and apply it to an existing selective desensitization neural network. The result of the evaluation experiment under realistic conditions indicates that the proposed method is rather better than the conventional methods in terms of correct and false detections. The proposed method also does not require more computational time with an increase in the number of classes, complicated parameter setting, and kernel design, which provides a highly practical method for EMG pattern classification.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Selective Desensitization / Neural Network / Electromyogram Signals / Multi-Class Classification / False Detection
Paper # NC2012-3
Date of Issue

Conference Information
Committee NC
Conference Date 2012/6/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Multi-Class EMG Pattern Classfication using a Selective Desensitization Neural Network with a Half-vs-Half Method
Sub Title (in English)
Keyword(1) Selective Desensitization
Keyword(2) Neural Network
Keyword(3) Electromyogram Signals
Keyword(4) Multi-Class Classification
Keyword(5) False Detection
1st Author's Name Kazumasa HORIE
1st Author's Affiliation Graduate School of Systems and Infomation Engineering, University of Tsukuba()
2nd Author's Name Atsuo SUEMITSU
2nd Author's Affiliation Japan Advanced Insititute of Science and Technology
3rd Author's Name Masahiko MORITA
3rd Author's Affiliation Graduate School of Systems and Infomation Engineering, University of Tsukuba
Date 2012-06-28
Paper # NC2012-3
Volume (vol) vol.112
Number (no) 108
Page pp.pp.-
#Pages 6
Date of Issue