Hybrid RNN(Long Short-Term Memory) with CNN (Densenet201) model using EEG signals in detecting epileptic seizure in a human being

Abstract

According to the researchers, the population has gradually increased, and the WHO - World Health Organization (WHO) released that epilepsy affects about 65 million people. Various investigations have been conducted in order to identify epilepsy using electroencephalogram - EEG signals that help to record the patient details that have the capability to collect the signals received from electrodes. These signals are presented with the spatial distribution of specific fields focused on the brain. In the current research, we have utilized a hybrid technique with the integration of a Dense Convolutional Network (DenseNet201) and LSTM - Long Short-Term Memory for epileptic seizure identification utilizing EEG data to choose appropriate features utilizing WOA - Whale Optimization Algorithm and PSO. The main focus of the current study is to reduce the time-consuming process during the detection of disease and enhance the accuracy of the prediction that helps physicians to start the treatment earlier as possible.