Enhancing Deep Learning-Based Epileptic Seizure Detection with Generative AI Techniques
Résumé
Epilepsy is a neurological disorder affecting millions of people worldwide. Early detection of seizures is critical for timely intervention. In this paper, we propose a novel deep learning approach using a **Bidirectional Long Short-Term Memory (BLSTM)** network combined with a **Variational Autoencoder (VAE)** and **Continuous Wavelet Transform (CWT)** for enhanced seizure detection from EEG signals. Our model outperforms existing models in terms of sensitivity, specificity, and generalization across different patients. Additionally, data augmentation through VAE addresses the class imbalance issue, significantly improving seizure detection performance. The results demonstrate a significant improvement in sensitivity to 96.3%, specificity to 99%, and G-Mean to 98.1% when evaluated on the CHB-MIT dataset.
Origine | Fichiers produits par l'(les) auteur(s) |
---|