Segmentation of Atrial Electrical Activity in Intracardiac Electrograms (IECGs) Using Convolutional Neural Network (CNN) Trained on Small Imbalanced Dataset

Hejc, Jakub; Pospíšil, David; Novotná, Petra; Pešl, Martin; Janoušek, Oto; Ronzhina, Marina; Stárek, Zdeněk

doi:10.23919/cinc53138.2021.9662729

Cited by 1 publication

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…in case of AV nodal re-entry tachycardia). The presented database can be used for developing the methods for signal processing analysis, as was previously presented in [12]. In the future, we expect including more subjects in the database.…”

Section: Discussionmentioning

confidence: 91%

Arrhythmia Database with Annotated Intracardial Atrial Signals from Pediatric Patients Undergoing Catheter Ablation

Ředina¹,

Hejc²,

Pospíšil³

et al. 2022

Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

Conventional surface ECG analysis has brought a breakthrough in the diagnosis of heart diseases in the last century. With improvements in technology, doctors are now able to read ECGs using catheters placed directly in the heart. Signals obtained in this way may herald a further shift in diagnostic capabilities. Early detection of cardiac arrhythmias is still one of the challenges in medicine. For this task, intracardiac ECG recordings from presented database can be a valuable source of information. The recordings were obtained during electrophysiological procedures on paediatric patients with arrhythmias.The database consists of a total of 326 records, from one hundred patients. For each record, global annotation with arrhythmia findings and local annotations with atrial activity onset/offset are available. Annotations capturing atrial activity were manually created by the experienced electrophysiologists. A large proportion of the records (n = 191) contain only sinus rhythm. Ventricular preexcitation (n = 58) and atrial premature beat (n = 47) were the most common findings in the database. We hope that the database presented in the paper will open new possibilities in the research of cardiac arrhythmias as well as in the improvement of wave detection in ECG or in the overall ECG delineation.

show abstract

Section: Discussionmentioning

confidence: 91%