Title :
Clustering analysis and recognition of the EMGs
Author :
Ling, Huang ; Bo, You ; Lina, Zhou
Author_Institution :
Coll. of Autom., Harbin Univ. of Sci. & Technol., Harbin, China
Abstract :
In order to identify EMGs better, the separability and clustering is compared for different features of EMGs. Then the six channels EMGs from forearm are identified based on the features with better separability. The EMGs of 18 motions of a hand are collected, the time domain features and the frequency domain features of the motions are extracted, then the separability and clustering of the features are analysized, in the end the time domain features are sent to three classifiers, which are built for the thumb, forefinger and the other three fingers, for identification. The accuracy of distinguishing is 98%, 97% and 100% respectively.
Keywords :
backpropagation; electromyography; feature extraction; gesture recognition; medical signal processing; neural nets; pattern clustering; source separation; time-frequency analysis; EMG recognition; backpropagation neural network; feature clustering; feature extraction; feature separability; forefinger; frequency domain features; gesture recognition; thumb; time domain features; Accuracy; Electromyography; Feature extraction; Fingers; Frequency domain analysis; Support vector machines; Time domain analysis;
Conference_Titel :
Intelligent Control and Information Processing (ICICIP), 2011 2nd International Conference on
Conference_Location :
Harbin
Print_ISBN :
978-1-4577-0813-8
DOI :
10.1109/ICICIP.2011.6008240
