View Item 

A multiclassifier system with dynamic ensemble selection applied to the recognition of EMG signals for the control of bio-prosthetic hand

Access Status
Fulltext not available
Authors
Kurzynski, M.
Woloszynski, Tomasz
Wolczowski, A.
Date
2010
Type
Conference Paper

Metadata
Show full item record
Citation
Kurzynski, M. and Woloszynski, T. and Wolczowski, A. 2010. A multiclassifier system with dynamic ensemble selection applied to the recognition of EMG signals for the control of bio-prosthetic hand.
Source Title
2010 3rd International Symposium on Applied Sciences in Biomedical and Communication Technologies, ISABEL 2010
DOI
ISBN
School
Department of Mechanical Engineering
URI
Collection
Abstract

The paper presents a concept of hand movements recognition on the basis of EMG signal analysis. Signal features are represented by coefficient of autoregressive (AR) model, and as classifier the original multiclassifier systems with dynamic ensemble selection are applied. The performance of the proposed methods was experimentally compared against three classifiers using real datasets. The systems developed achieved the highest overall classification accuracies demonstrating the potential of dynamic classifier selection for recognition of EMG signals. ©2010 IEEE.