BUS, M. Gabriel (2024) Application de la décomposition en modes par noyaux reproduisants (KMD) à la détection de crises épileptiques dans les électroencéphalogrammes (EEG) PRE - Research Project, ENSTA.

[img]
Preview
PDF
1697Kb

Abstract

This project aims to explore the application of the Kernel Mean Decomposition (KMD) method for detecting epileptic seizures from intracranial electroencephalogram (iEEG) signals. The objective is to enhance this method by using both simulated signals, generated via stochastic differential equations, and real signals from a public database. The KMD method, as an alternative to machine learning, has the potential to identify seizure patterns, whose criteria are not always clearly defined. With a large set of EEG data, the effectiveness of this method will be assessed and optimized using data science tools. The KMD method will be tested on simulated signals, derived from stochastic differential equations, and then on real signals from medical databases, notably from ETH Zurich. We will compare the results obtained by this method to a machine learning method developed by researchers at ETH Zurich.

Item Type:Thesis (PRE - Research Project)
Uncontrolled Keywords:Kernel Mode Decomposition (KMD), electroencephalogra- phy (EEG), epileptic seizure detection.
Subjects:Mathematics and Applications
ID Code:10165
Deposited By:Gabriel BUS
Deposited On:03 sept. 2024 10:33
Dernière modification:03 sept. 2024 10:33

Repository Staff Only: item control page