Statistical feature rich Deep learning based Epileptic Seizure detection

Epilepsy is a neurological condition characterised by abnormal behaviour and recurring seizures as a result of abnormal brain activity. Epilepsy diagnosis necessitates the examination of a huge amount of EEG data, which is both time-consuming and complex. Epilepsy is classified as unusual spikes in the EEG recordings. This research demonstrates utilizing statistical feature extraction followed by deep learning techniques LSTM and CNN to diagnose epilepsy from multi-channel EEG recordings. The method was created to address the shortcomings of previous Machine Learning methods that fail when dealing with massive and multi-channel datasets. Much importance has been placed in explaining the significance of the statistical features that are used as feature extraction. Furthermore, the proposed method has been validated on the clinical CHB-MIT dataset, providing validation accuracy of 94.5% and specificity of 96.3%.