Evaluation Of ANN And CNN In Gesture Classification For Myoelectric Prosthesis Hand

Lee, Su Wing (2016) Evaluation Of ANN And CNN In Gesture Classification For Myoelectric Prosthesis Hand. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Evaluation Of ANN And CNN In Gesture Classification For Myoelectric Prosthesis Hand.pdf - Submitted Version Download (247kB)

Abstract

The numbers of limb amputees are increasing yearly due to the wars, trauma or illness and they are suffering from different types of pain and reduced health-related quality of life. The existing prosthetic allows an amputee to partially restore the ability and function of the lost limb. However the prosthetic available in the market mostly concentrated on the cosmetic purpose and robotic prosthesis hand are too expensive to be afforded by most amputees. To restore the hand ability, myoelectric signal played a significant role in pattern recognition for gesture classification. In this project, in-depth study in various feature extraction and the classifier include ANN and CNN is conducted to get the high accuracy of gesture classification. By integrating the surface electromyography sensors, the myoelectric signal can be recording from the remaining limb. A pattern recognition algorithm will be developed to identify the motion desired by the amputee and control the actuators accordingly. Finally with the convolutional neural network, it is expected that this project will able to deliver a high accuracy of gesture classification.

Actions (login required)

View Item