Exploring CNN Architectures for Seizure Detection from MRI during Pregnancy

Deep Learning (DL) represents a subset that has revolutionized traditional machine learning. Unlike earlier methods relying on handcrafted feature extraction, DL automates both feature extraction and classification. This innovation has significantly advanced applications in various medical domains, including the diagnosis of epileptic seizures. This research provides a comparative overview of studies concentrating on automated epileptic seizure detection using leading convolutional neural networks (CNN) architectures. It explores diverse architectures proposed for the automatic diagnosis of epileptic seizures through MRI, encompassing an analysis of rehabilitation systems utilizing DL. The proposed work investigates the application of four primitive CNN architectures on the identification of onset of seizure. The proposed architectures are validated through benchmark MRI samples. Experimental evaluation is made on samples of pregnant females. Overall rate of accuracy stands at 85%, 80%, 90%, and 95% respectively.