A machine learning-based system to detect precursor symptoms of epileptic seizures using EEG data in real time.
This project aims to develop a real-time seizure precursor detection system using EEG signals.
By leveraging deep learning models such as CNN and Bi-LSTM, the system predicts the likelihood of epileptic seizures based on EEG waveforms.
- ✅ Deep Learning Architecture: Combined 1D CNN + Bi-LSTM model for temporal EEG signal analysis
- ✅ Real-time EEG Estimator: Efficient, lightweight, and compatible with mobile devices
- ✅ Seizure Alert System: Provides early warnings to help prevent accidents (e.g., falls, head trauma)
The trained model used for inference (
model_fold1_best.pt) is available at the link below:
🔗 Download model_fold1_best.pt from Google Drive
Please make sure to place it in the following directory before running the server:
/backend_server/saved_models/model_fold1_best.pt
To run the backend server:
cd backend_server
uvicorn main:app --reload