18. Akademik Bilisim Konferansi

BaşlıkProtez-Biyonik El Kontrolü İçin EMG İşaretlerinin Makine Öğrenmesi Metodlarıyla Sınıflandırılması
ÖğrenciHayır
Yazar(lar) Yazar 1
Name: Duygu BAĞCI
Org: Dokuz Eylül University
Country: TR
E-mail: duygu.bagci_AT_deu.edu.tr

Yazar 2
Name: Osman Hilmi KOÇAL
Org: Yalova University
Country: TR
E-mail: osman.kocal_AT_yalova.edu.tr
Anahtar KelimelerEMG, Machine Learning, Signal Classification, Markov Models, Wavelet Transform.
ÖzetSubmission Title: Classification of EMG Signals with Machine Learning Methods to Control Prosthetic-Bionic Hand Abstract: The aim of this paper is to present the results of study about classification of electromyography (EMG) signals collected from forearm muscles to control prosthetic-bionic hand. Dataset used in this study are consist of 900 different patterns on six different hand movements. Each pattern for specific movement obtained with queuing of EMG signals simultaneously recorded with electrodes. Markov Model and Wavelet Transform used for acquire attribute vectors from raw EMG signals. Different machine learning algorithms applied to attribute vectors and its classification performances are examined and compared. Average classification accuracy obtained with K-nearest neighbor based IBK algorithm is %93,78.
Başlıklar AB-Bildiri
Veri Madenciliği
Dosya  
 

 

Powered by OpenConf®
Copyright ©2002-2015 Zakon Group LLC