Poster abstract: Detecting epileptic seizures with a smartphone using frequency analysis
| Autor: | Ahmed Helmy, Amir Helmy |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Computer science
05 social sciences Fast Fourier transform Real-time computing Mobile computing Wearable computer medicine.disease 050105 experimental psychology 03 medical and health sciences Statistical classification Epilepsy 0302 clinical medicine Robustness (computer science) medicine 0501 psychology and cognitive sciences Android (operating system) Cluster analysis 030217 neurology & neurosurgery |
| Zdroj: | INFOCOM Workshops |
| DOI: | 10.1109/infcomw.2018.8406999 |
| Popis: | Seizures affect many millions around the world, and cause injury and death if not treated promptly. Their efficient detection provides a major challenge to the mobile health, and pervasive/wearable computing communities, due to the varying scenarios of convulsions and movement involved. In this study, we introduce a frequency-analysis based algorithm, using a novel metric, for detection of major and minor convulsions in generalized and focal seizures. Our algorithm, using FFT analysis, is implemented and tested extensively using our Seizario app for Android with promising results. Our experiments provide the optimal detection strategy using a spectral contribution metric. With the proposed strategy, robust classification can be obtained with a margin over 15%, and variation of less than 3%, providing very high-quality clustering of activities. This enables smart networking with care-givers to respond to epileptic emergencies in a timely manner. This work provides, for the first time, efficient and practical detection, and notification of seizure-related emergencies via pervasive mobile computing. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|