Epileptic Seizure Prediction Based On Machine learning And Convolution Neural Network

Epileptic Seizure Prediction Based On Machine learning And Convolution Neural Network.

Presented by Omar Sayed Dosouky Ahmed Nasser Abo elwafa Aya Abdullah Farag Shorouk Osama Mohamed.

epilepsy. Seizures are bursts of electrical activity in the brain that temporarily affect how it works that can cause a wide range of symptoms, they can start at any age, but usually starts either in childhood or in people over 60.

Epilepsy. It affects 50 million people around the world, according to the World Health Organization (WHO).

Solution. A seizure prediction system for warning the caregivers of the upcoming seizure to give the patient the required medication. to reach the ultimate goal which is improving the patient’s quality of life..

EEG Signal states. inter-ictal pre-ictal. Interictal State Seizure o Seizure Pre ct on Pre-lctal State prediction Horizon (1 Ictal State Seizure Onset hr Postictal State Time.

How can we predict seizures?.

seizure prediction. Seizure prediction can happen in the event of differentiation between the pre-ictal and the ictal states..

Cloud storage and processing EEG headset Data acquisition and transmission cec WiFi 3G/4G Pre-processing & data management Feature extraction Feature selection Feature classification Database Patient Bluetooth Smartphone GPS-based alert system Family member Hospital.

Proposed Methods.

Machine L earning Convolution Neural Network Mobile Application.

Machine learning.

k nearest neighbor Random Forest Support vector machine.

Accuracy 95.3% sensitivity 99.1% specificity 92.5%.

Convolution Neural Network.

We applied a 4-layer CNN architecture on spectrograms of the signals.

Mobile Application.

Flutter Firebase Flask. Tools.

Feel free to approach us if you have any questions. Thank you!.