Abstract
Seizures affect a significant portion of the world's population and while a small proportion of the seizures are easy to detect, the vast majority are subtle enough to require the expertise of a neurologist. In this paper, we present a multifaceted computational approach to detecting the presence and locality of a seizure using Electroencephalography (EEG) signals. We test the proposed approach using a variety of signals and demonstrate the efficacy of the approach.