Multi-task Deep Neural Network for Intracardiac Activity Reconstruction in Atrial Fibrillation
Year: 2024
Type: conference-paper
Abstract: The reconstruction of cardiac mapping for arrhythmia characterization remains a challenge for both invasive and non-invasive approaches.This paper presents a novel data-driven approach using a multi-task deep neural network for the reconstruction of intracardiac activity from body surface potentials in atrial fibrillation patients.The proposed framework integrates 3D convolutions and Recurrent Neural Networks for spatio-temporal feature extraction from body surface potentials, enhancing the reconstruction of electrograms.Performance was evaluated against the Tikhonov method in synthetic and real data, showing robustness to noise and electrode disconnection.
Source: Computing in cardiology
Language: English
Topic: ECG Monitoring and Analysis
Subfield: Cardiology and Cardiovascular Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W4405435666