Comparative Study : Epileptic Seizure Prediction Systems (Unification of multiple systems to improve the efficiency)

Abstract

With the advent of technology, from the modes of communication to disease diagnosis, all the fields went through a drastic change. Epileptic seizure is also one of those neurological conditions whose predictability methods are being researched by e-health researchers in order to reach an accurate prediction method. An epileptic seizure is not fatal but can prove fatal if it occurs under certain circumstances that can put patient’s life at risk. Therefore, prediction methods can help the caretaker or the emergency services early and allow precautions to be taken accordingly. This thesis is a comparative study of different epileptic prediction methods which have been provided by various researchers and have available literature. The features are being extracted from body vitals EEG, ECG, and an accelerometer. The purpose of implementing the feature extraction from different research papers was to achieve higher accuracy utilising available methods in the literature. In case of a seizure, an alarm will be triggered to alert the patient along with alerting the caregiver(s). Moreover, the patient’s data will be stored in a cloud to help the doctor with a better diagnosis. The unification of these methods envisions a system that can provide higher accuracy and lead to a better life quality for epileptic patients.