Article,
An investigation of the contextual distribution of false positives in a deep learning-based atrial fibrillation detection algorithm
Contributors
DOMINGUEZ, Helena
0000-0002-7089-2636
[2]
[3]
Sharma, Kamal
0000-0001-5866-2566
[4]
Bardram, Jakob
0000-0003-1390-8758
[1]
Affiliations
- [1] Tech Univ Denmark, Dept Hlth & Technol, DK-2800 Lyngby, Denmark [NORA names: DTU Technical University of Denmark; University; Denmark; Europe, EU; Nordic; OECD];
- [2] Bispebjerg & Frederiksberg Hosp, Dept Cardiol, DK-2400 Copenhagen, Denmark [NORA names: Capital Region of Denmark; Hospital; Denmark; Europe, EU; Nordic; OECD];
- [3] Bispebjerg & Frederiksberg Hosp, Dept Cardiol, DK-2400 Copenhagen, Denmark [NORA names: KU University of Copenhagen; University; Denmark; Europe, EU; Nordic; OECD];
- [4] UN Mehta Inst Cardiol & Res Ctr, Civil Hosp Campus, Ahmadabad 380016, Gujarat, India [NORA names: India; Asia, South]
Abstract
Goal: To investigate the contextual and temporal distribution of false positives (FPs) in a state-of-the-art deep learning (DL)-based atrial fibrillation (AF) detection algorithm when applied to an electrocardiogram (ECG) dataset collected under free-living ambulatory conditions. We hypothesize that under such conditions, the FPs detected by a DL model might have some correlations with the patient's ambulatory contexts.
Keywords
Arrhythmias,
Atrial fibrillation (AF),
Context-aware ECG,
Deep learning (DL),
Electrocardiogram (ECG),
False positive (FP)
