Record Details

Comparison of Different Time and Frequency Domain Feature Extraction Methods on Elbow Gesture’s EMG

European Journal of Interdisciplinary Studies

View Archive Info
 
 
Field Value
 
Title Comparison of Different Time and Frequency Domain Feature Extraction Methods on Elbow Gesture’s EMG
 
Creator Altın, Cemil
Er, Orhan
 
Description Objective:In this study we will get EMG signals from arm for different elbow gestures, than filtering the signal and later classification the signal. The reason for doing is that, EMG signals are used for many rehabilitation and assistive prostheses of paralyzed or injured people. Methods:Filtering a biological signal is the key point for these type studies. Filtering the EMG signals needed and starts with the elimination of the 50 Hz mains supply noise. After filtering the signal, feature extraction will be applied for both wrist flexion and wrist extension cases. There are many feature extraction methods for time and frequency domain. After feature extraction, classification of hand movements will be studied using extracted features. Classification is made using K Nearest Neighbor algorithm. The dataset used in this study is acquired by the EMG signal acquisition tool and belong to us. Results:90 % accuracy performance is obtained by K Nearest Neighbor algorithm purposed signal classification. Conclusion:This system is capable of conducting the classification process with a good performance to biomedical studies. So,this structure can be helpful as machine-learning based decision support system for medical purpose.
 
Publisher EUSER
 
Date 2016-08-30
 
Type info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
Peer-reviewed Article
 
Identifier http://journals.euser.org/index.php/ejis/article/view/1884
10.26417/ejis.v2i3-35-44
 
Source European Journal of Interdisciplinary Studies; Vol 2 No 3 (2016): May-August 2016; 35-44
2411-4138
2411-958X
 
Language eng
 
Relation http://journals.euser.org/index.php/ejis/article/view/1884/1866