Epilepsy Seizure Identification Using the One Dimensional Convolutional Neural Network from Electroencephalogram Signals

Epilepsy is a chronic, noncommunicable brain illness affecting 50 million people worldwide. It has a significant impact on the persons who are afflicted, both in terms of their condition financially and the quality of life. The present scarcity of neurologists proficient in diagnosing epilepsy, coupled with significant differences in outcomes and accessibility, underscores the heightened need of accurate epileptic interpretation. Furthermore, the constraints that are now present in clinical and conventional machine learning diagnostic approaches give a reason for the need of developing an automated system that makes use of a deep learning model to identify and monitor epilepsy by making use of a large database. In this work, a new real-time electroencephalogram (EEG) -based approach known as AFAWT₁DCNN is presented for the purpose of identifying epileptic seizures. This technique makes use of a 1D-convolutional neural network (1D-CNN) in addition to the Adaptive Flexible Analytic Wavelet Transform. This research aims to identify the most distinguishing channel and examine the optimal number of characteristics necessary for optimal system performance. The article also presents the methodology for elucidating the individual and aggregate predictions of the classification model. The proposed technique's dependability and accuracy may be related to the evaluation of the model's performance using ten-fold cross-validation. Unlike the present scarcity of deep models used in epilepsy diagnostic research investigations, our proposed technique is more straightforward and demands fewer computational resources.