18th National Biomedical Engineering Meeting (BIYOMUT), İstanbul, Turkey, 16 - 17 October 2014
In this study, a structure based ona Self Organizing Map (SOM) depending on RMS(Root Mean Square), MAV(Mean Absolute Value) and MF(Mean Frequency) features was formed in recording the EMG(Elektromyogram) signals during the performof 24 different movements in hand and fingers to detect of active electrodes. Recorded data with surface EMG electrodes, from 24 channels with 2 kHz sampling frequency as bipolar primarily ispreprocessed. In preprocessing, these data were filtered with 50 Hz notch filter, 3-450 Hz frequency band was selected using the 6th order Butterworth band-pass filter. RMS, MAV and MF features extracting from this EMG data were defined as SOM classifier input. Then, active channels in the classifier output were found for each features and resultswere compared with each other.