EMG- BASED HAND GESTURE RECOGNITION USING DEEP LEARNING AND SIGNAL-TO-IMAGE CONVERSION TOOLS
Main Article Content
Abstract
In this paper, deep learning-based hand gesture recognition using surface EMG signals is presented. We use Principal component analysis (PCA) to reduce the data set. Here a threshold-based approach is also proposed to select the principal components (PCs). Then the Continuous wavelet transform (CWT) is carried out to prepare the time-frequency representation of images which is used as the input of the classifier. A very deep convolutional neural network (CNN) is proposed as the gesture classifier. The classifier is trained on 10-fold cross-validation framework and we achieve average recognition accuracy of 99.44%, sensitivity of 97.78% and specificity of 99.68% respectively.
Article Details
How to Cite
AKTER, S., PRAMANIK, B. K., & HAMID, M. E. (2023). EMG- BASED HAND GESTURE RECOGNITION USING DEEP LEARNING AND SIGNAL-TO-IMAGE CONVERSION TOOLS. Journal of Engineering Studies and Research, 29(2), 7-19. Retrieved from https://jesr.ub.ro/index.php/1/article/view/375
Section
Articles