Related Experiment Video
Updated: Dec 20, 2025
Estimating Bilateral Atrial Function by Cardiovascular Magnetic Resonance Feature Tracking in Patients with Paroxysmal Atrial Fibrillation
Published on: July 20, 2022
MultiFusionNet: Atrial Fibrillation Detection With Deep Neural Networks
Luan Tran1, Yanfang Li1, Luciano Nocera1
1University of Southern California, Los Angeles, CA, USA.
Insights
MultiFusionNet accurately classifies atrial fibrillation (AF) using a novel deep learning approach. This method fuses extracted ECG features and raw data, outperforming existing algorithms for reliable arrhythmia detection.
Area of Science:
- Cardiology
- Artificial Intelligence
- Biomedical Engineering
Background:
- Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia, posing significant risks for heart failure and coronary artery disease.
- Detecting AF via short electrocardiogram (ECG) recordings is crucial but challenging due to noise and similar rhythms.
- Accurate discrimination of AF from normal sinus rhythm and other arrhythmias requires advanced analytical methods.
Purpose of the Study:
- To develop and evaluate MultiFusionNet, a deep learning network for accurate atrial fibrillation classification from short ECG recordings.
- To investigate the efficacy of a multiplicative fusion method combining extracted features and raw ECG data.
- To compare the performance of MultiFusionNet against existing algorithms that utilize features or raw data independently.
Main Methods:
- Proposed MultiFusionNet, a deep learning architecture employing multiplicative fusion of two networks.
- Trained sub-networks on distinct data sources: extracted ECG features and raw ECG data.
- Experimentally validated the classification accuracy and performance against state-of-the-art methods.
Main Results:
- MultiFusionNet achieved superior accuracy in classifying atrial fibrillation compared to methods using only extracted features or raw data.
- The multiplicative fusion strategy significantly enhanced the model's ability to leverage both knowledge sources.
- The proposed fusion method outperformed other combination techniques evaluated in the study.
Conclusions:
- MultiFusionNet offers a highly accurate and effective deep learning solution for atrial fibrillation detection from short ECGs.
- Combining extracted features and raw data through multiplicative fusion is a promising strategy for improving arrhythmia classification.
- This approach holds potential for enhancing diagnostic tools in clinical cardiology and digital health applications.
Abstract:
Atrial fibrillation (AF) is the most common cardiac arrhythmia as well as a significant risk factor in heart failure and coronary artery disease. AF can be detected by using a short ECG recording. However, discriminating atrial fibrillation from normal sinus rhythm, other arrhythmia and strong noise, given a short ECG recording, is challenging. Towards this end, we propose MultiFusionNet, a deep learning network that uses a multiplicative fusion method to combine two deep neural networks trained on different sources of knowledge, i.e., extracted features and raw data. Thus, MultiFusionNet can exploit the relevant extracted features to improve upon the utilization of the deep learning model on the raw data. Our experiments show that this approach offers the most accurate AF classification and outperforms recently published algorithms that either use extracted features or raw data separately. Finally, we show that our multiplicative fusion method for combining the two sub-networks outperforms several other combining methods.
