Mr Yusuf Dinah

Research project: Developing Real-Time Seizure Detection Using Deep Learning Architectures

Abstract

Epilepsy is a chronic illness of the brain whereby the patient cannot control the random firing of the brain's electrical activity, resulting in frequent unpredictable seizures. This disease affects over 50 million people worldwide (Ferreira, J. et al., 2009). It's the monitoring of the seizures that lead to the diagnosis of Epilepsy (Thomas et al., 2018).

The aim of the project is to design, develop and evaluate novel strategies to enhance the current methods of detecting seizures presented by epilepsy patients, using EEG data. Successful completion of the project involves developing novel state of the art architectures, in the following manner: 

• A new 2D and CNN Bi-LSTM Architecture for EEG Analysis
• A Legendre Memory Units-based Architecture for Real-time EEG Analysis
• A Hybrid LSTM with Legendre Memory Units for Real-time EEG Analysis