Personalization of Cardiac Electrophysiology Model using the Unscented Kalman Filtering

Abstract : Cardiac electrophysiology mapping techniques now allow to record denser intra-operative electrograms (ECG). The patient-specific information extracted from these clinical recordings is extremely valuable. A growing field of research focuses on the personalization of electro-physiology models using this patient-specific information. The modeling in silico of a patient electrophysiology is needed to better understand the mechanism of cardiac arrhythmia. In the scope of ischemic cardiomyopa-thy, the predictive power of patient-specific simulations may also provide a substantial guidance in defining the optimal location of the implantable defibrillator, since all possible configurations could be tested in silico. This article describes an innovative personalization approach based on an unscented Kalman filter. Following an iterative process, the apparent conductivity is efficiently estimated in specific regions. A sensitivity analysis is performed to assess the filter parameters. With three patient cases, we finally demonstrate the accuracy and efficiency of our algorithm.
Type de document :
Communication dans un congrès
Computer Assisted Radiology and Surgery (CARS 2015), Jun 2015, Barcelona, Spain
Liste complète des métadonnées

Littérature citée [14 références]  Voir  Masquer  Télécharger


https://hal.inria.fr/hal-01195719
Contributeur : Hugo Talbot <>
Soumis le : mardi 10 novembre 2015 - 10:11:37
Dernière modification le : jeudi 30 novembre 2017 - 09:21:10
Document(s) archivé(s) le : vendredi 12 février 2016 - 17:17:12

Fichiers

CARS2015-HTalbot.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01195719, version 1

Collections

Citation

Hugo Talbot, Stephane Cotin, Reza Razavi, Christopher Rinaldi, Hervé Delingette. Personalization of Cardiac Electrophysiology Model using the Unscented Kalman Filtering. Computer Assisted Radiology and Surgery (CARS 2015), Jun 2015, Barcelona, Spain. 〈hal-01195719〉

Partager

Métriques

Consultations de la notice

318

Téléchargements de fichiers

360