Convolutional Neural Network (CNN) implementation on bionic hand in comparison with Recurrent Neural Network (RNN)
The main objective of this thesis is to improve the implementation of gesture classifier on Bionic Hand, which uses the Electromyography (EMG) signal from the Myo Armband device. The improvement is mainly on the accuracy and the additional number of gestures to be added. It utilized the Convolutional Neural Network (CNN) machine learning as the main gesture classifier, and then was compared with Recurrent Neural Network (RNN) machine learning. This thesis also includes the development and modification of the program script, along with the analysis of each results. At the end, the CNN was able to achieve an accuracy of 90.32%, while the RNN gesture classifier could reach higher accuracy at 96.32%. Besides that, the pointing gesture and rock gesture are successfully implemented and classified by both methods.
|B02966||(Rack Thesis)||Available but not for loan - Missing|
No other version available