An embedded system for automatic pathological characterization classification by EEG

This study presents an system for the automatic classification of pathological EEG signals using artificial intelligence (AI) algorithms. The research explores the application of deep learning and machine learning techniques, including 1D CNN and KNN to classify EEG signals into different pathological states. The system utilizes the Bonn University EEG dataset, which includes normal, pre-ictal, and seizure EEG recordings. A wavelet transform is employed to extract EEG features, followed by classification using KNN. And raw data for CNN, achieving high accuracy (98.9% for KNN and 97.78% for CNN). The experiments are conducted in a PC-based environment, utilizing an Intel Core i7 and an NVIDIA GTX 1650 GPU for training and evaluation. This work contributes to the field of biomedical signal processing and neurological disorder detection, offering a promising solution for the automated diagnosis of epilepsy. Future work includes expanding the dataset, optimizing computational efficiency, and deploying the system in clinical environments.