Details
Presenter(s)
![Farnaz Forooghifar Headshot](https://confcats-catavault.s3.amazonaws.com/CATAVault/ieeecass/master/files/styles/cc_user_photo/s3/user-pictures/10611.jpg?h=bb777b72&itok=S_I4080_)
Display Name
Farnaz Forooghifar
- Affiliation
-
AffiliationEmbedded Systems Laboratory / École Polytechnique Fédérale de Lausanne
- Country
-
CountrySwitzerland
Abstract
Low-power wearable technologies offer a promising solution to pervasive epilepsy monitoring by removing time and location constraints, and fulfilling long-term tracking. Since epileptic seizures occur infrequently, an anomaly detection approach reduces the amount of data needed before being able to detect the next coming seizures. This work combines the concepts of self-aware system and anomaly detection to provide an energy-efficient, general solution for seizure detection, which is personalized after analyzing the first seizure of each patient. The system, then, uses a simple anomaly detection model whenever the decision is self-assessed as reliable, and relies on a more complex model otherwise.