Improved detection of atrial fibrillation in high-resolution long-term electrocardiography
based on asynchronous recording

Atrial fibrillation (AF) is frequent and may have serious consequences like strokes. Long-term ECG recording is routinely performed to diagnose AF. Unfortunately, contemporary software tools show a low accuracy to automatically detect AF.

Fiche signalétique

  • Département HESB | Technique et informatique
  • Pôle de recherche Human Centered Engineering
  • Direction du projet Thomas Niederhauser

Contexte initial

Atrial fibrillation (AF) is frequent and may have serious consequences like strokes. Long-term ECG recording is routinely performed to diagnose AF. Unfortunately, contemporary software tools show a low accuracy to automatically detect AF. The project aims to improve AF detection based on level-crossing sampling. The acquired, non-equidistant ECG signals feature more precise time information which enables to develop a model-based, rhythm classification system.