A Robust Detection and Classification of Epileptic Seizures using Deep Learning Approach

The brain, one of the most vital organs in the human body, is responsible for the functions of the entire nervous system. Conditions affecting the brain pose significant risks to overall health. Epilepsy, along with other neurological conditions such as Parkinson’s disease (PD) and Alzheimer’s disease (AD), is characterised by uncontrollable surges of neurochemical activity in the brain, which ultimately result in seizures. Identifying these seizures is crucial for successful research and treatment. Electroencephalography (EEG) is a technique commonly used to analyze brain activity and diagnose neurological disorders like epilepsy. Historically, neurologists have dedicated considerable time to visually inspecting EEG data to identify anomalies associated with epilepsy. This work introduces a deep learning (DL) strategy for categorizing epileptic episodes based on its findings. To implement DLtechniques like Convolutional Neural Networks (ConvNet), we utilized a dataset from Bonn University. This dataset contains spectral data from 128 channels. Seizure classification was specifically done using the VGG19 DL method. The ConvNet model achieves an outstanding level of accurate categorization with a testing classification accuracy range of 97.82% to 99.57%.