Development of computer aided design for EEG signals epilepsy diagnosis using artifical neural network

Ahmezul, Ahmezan (2021) Development of computer aided design for EEG signals epilepsy diagnosis using artifical neural network. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Development of computer aided design for EEG signals epilepsy diagnosis using artifical neural network.pdf - Submitted Version Download (537kB)
	Text (Full text) Development of computer aided design for EEG signals epilepsy diagnosis using artifical neural network.pdf - Submitted Version Restricted to Registered users only Download (3MB)

Abstract

Epilepsy is a brain condition that affects the whole brain nervous system and is characterised by high-frequency and high-voltage brain waves called seizures. This disorder is identified as one of the uncontrolled movements shown by epilepsy patients during an outbreak, resulting in loss of consciousness and convulsions. As a result, the purpose of this thesis is to construct an EEG Epilepsy Recognition System using Artificial Neural Networks (ANN). Their principal tool is the Cascade-Forward Neural Network technology, which their system designed to perform a process similar to that of a human brain. This brain-inspired technology was designed to mimic how human brains think. This thesis offers an epilepsy detection process implemented in MATLAB utilising Cascade-forward Neural Networks. Additionally, this study employed the Electroencephalogram (EEG) signal to diagnose and access human brain activity and disturbance by using a dataset collected from the University of Bonn (Bonn), which has been extensively used by other researchers doing epilepsy research. The MindLink EEG Sensor is used to collect external EEG data, which is subsequently utilised to test the neural network. As for the result, this Artificial Neural Network successfully carried out with 77.1% for training, 77.3% for validation, 74.7% for testing and lastly the overall accuracy is 76.2% by using 15 hidden neuron network.

Actions (login required)

View Item