Hybrid CNN-GRU Model for Seizure Detection Using EEG Biomedical Signal Data

In the field of biomedicine, using electroencephalogram (EEG) statistics to locate epileptic seizures has grown to be an important area of research. The essential motive for that is that seizures are unpredictable and might have a significant effect on a person's health. By offering EEGxNet-CGRU, a unique hybrid deep learning structure that combines Convolutional Neural Networks (CNNs) and Gated Recurrent Units (GRUs) to enhance the automated prognosis of seizures using electroencephalogram (EEG) data. The convolutional neural network (CNN) component is used to extract robust spatial and frequency-domain functions from raw multichannel EEG recordings. The GRU component, however, is used to locate long-range temporal correlations that are part of sequential mental activity patterns. The version is proven to have an excessive degree of sensitivity and specificity in distinguishing between seizure states and states that do not contain seizures. To teach and examine the usage of a preferred EEG dataset. The consequences of an evaluation study showed that EEGxNet-CGRU is more effective at classifying information and uses significantly less computing energy than standalone CNNs, GRUs, and standard system mastery classifiers. Because of this, the approach demonstrated presents a framework that can grow and function with real-time data, making it possible to create innovative and non-invasive seizure tracking structures.

Hali tarjima qilinmagan