EMG-Signal-for-Gesture-Recognition ✋
Machine learning based EMG Signal analysis for Gesture Recognition
Content
1) Overview
-The paper we are discussing titled "Machine Learning-Based Hand Gesture Recognition via EMG Data"
by Zehra Karapinar Senturk and Melahat Sevgul Bakay focuses on a machine learning-based
approach for hand gesture recognition using EMG data
-The authors explore the use of machine learning algorithms to recognize and classify specific hand gestures
based on the patterns observed in the EMG signals.
-Our main application is virtual rehabilitation by combining gesture recognition with VR technology.
2) Code explaination
Requirements
sklearn == 1.2.2
numpy==1.22.4
pandas == 2.0
matplotlib == 3.6.0
scipy == 1.10.1
A)Preprocessing:
preprocess the EMG raw signal data by applaying Notch filter then Band pass filter.
Raw EMG signal |
Notch filtered signal |
Band pass filtered signal |
|---|---|---|
 |
 |
 |
B) Feature extraction & selection:
After we preprocess the EMG signal , we now do feature extraction to extract the following 12 features from each of the 8 channels:
integrated EMG IEMG |
simple square integrated SSI |
Root mean squared rms |
mean absolute value MAV |
Variance |
waveform length WL |
peak to peak ptp |
difference absolute mean value DAMV |
difference absolute standard deviation value DASDV |
Willison amplitude WAMP |
min |
max |
|---|
features |
after extraction |
|---|---|
 |
 |
Then , we apply feature selection techniques Kbest in our case
Feature selection |
Feature scores |
|---|---|
 |
 |
C) Classification:
we used the following three classifiers to construct our model:
classifier |
code |
|---|---|
Random forest |
 |
SVM |
 |
MLP neural network |
 |
3) Model evaluation
Performance comparsion
Classifier |
Confusion Matrix |
|---|---|
SVM |
 |
RF |
 |
MLP |
 |