{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T05:57:07Z","timestamp":1790575027031,"version":"4.1.0"},"reference-count":43,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T00:00:00Z","timestamp":1790553600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Cardiac arrhythmias often progress without clear symptoms and may remain undetected until severe complications occur. Continuous monitoring through consumer smartwatches offers a scalable and non-invasive opportunity for real-time rhythm screening in free-living conditions. However, wrist-based photoplethysmography (PPG) signals are highly susceptible to motion artifacts and inter-subject variability, limiting traditional signal-processing techniques. In this work, we compare classical machine learning and deep learning approaches for multi-class arrhythmia classification from real-life smartwatch PPG and accelerometer recordings collected in the Pulsewatch clinical trial. Three rhythm categories are considered: normal sinus rhythm (NSR), atrial fibrillation (AF), and premature atrial or ventricular contractions (PAC\/PVC).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>The evaluated approaches include a Random Forest classifier based on engineered temporal and morphological features, together with two end-to-end architectures, CNN\u2013GRU and CNN\u2013BiLSTM, designed to learn temporal patterns directly from raw two-channel smartwatch waveforms while addressing motion artifacts, sensor noise, and cross-subject generalization.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Under two-fold subject-independent cross-validation, the Random Forest achieved the highest average accuracy among the proposed methods (91.22%), together with the best macro precision (85.89%), macro F\n                      <jats:sub>1<\/jats:sub>\n                      score (84.17%), and macro ROC-AUC (97.03%). The CNN\u2013BiLSTM achieved the highest macro recall (87.37%), indicating stronger sensitivity to arrhythmic classes, while the CNN\u2013GRU remained competitive with a macro recall of 86.85% and a macro F\n                      <jats:sub>1<\/jats:sub>\n                      score of 83.13%.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>\n                      Compared with the Pulsewatch baselines reported by Han et al., the proposed methods improved the available macro-level metrics, with the Random Forest obtaining the strongest macro precision and macro F\n                      <jats:sub>1<\/jats:sub>\n                      score and the CNN\u2013BiLSTM the strongest macro recall, indicating improved class-balanced recognition of clinically relevant arrhythmic events. These results show that feature-based models remain highly competitive for smartwatch-based arrhythmia detection, achieving the strongest overall and macro F\n                      <jats:sub>1<\/jats:sub>\n                      performance among the proposed methods, while hybrid convolutional\u2013recurrent architectures provide increased sensitivity to minority rhythm classes. Deployment-oriented profiling under an identical single-core benchmark showed a mean end-to-end inference time of 33.08 ms for the Random Forest pipeline and 51.87 ms for the CNN\u2013BiLSTM model on a 30-s segment, supporting near-real-time rhythm screening.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>Overall, this work highlights the trade-offs between the analyzed approaches in terms of accuracy, class-balanced performance, computational complexity, and deployment feasibility for wearable cardiac monitoring systems.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2026.1882739","type":"journal-article","created":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T05:44:01Z","timestamp":1790574241000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Development of an efficient deep learning model for arrhythmia detection from consumer smartwatches"],"prefix":"10.3389","volume":"9","author":[{"given":"Dan-Cristian","family":"George","sequence":"first","affiliation":[{"name":"Department of Computer Science, Faculty of Mathematics and Computer Science, Babe\u015f-Bolyai University","place":["Cluj-Napoca, Romania"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mihai-Alexandru","family":"Hulubeac","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Faculty of Mathematics and Computer Science, Babe\u015f-Bolyai University","place":["Cluj-Napoca, Romania"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"\u015etefan-Lucian","family":"Gr\u0103mad\u0103","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Faculty of Mathematics and Computer Science, Babe\u015f-Bolyai University","place":["Cluj-Napoca, Romania"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zolt\u00e1n","family":"B\u00e1lint","sequence":"additional","affiliation":[{"name":"Department of Biomedical Physics, Faculty of Physics, Babe\u015f-Bolyai University","place":["Cluj-Napoca, Romania"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Laura","family":"Dio\u015fan","sequence":"additional","affiliation":[{"name":"Department of Computer Science, Faculty of Mathematics and Computer Science, Babe\u015f-Bolyai University","place":["Cluj-Napoca, Romania"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,28]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"2848","DOI":"10.3390\/s25092848","article-title":"Wearable devices for arrhythmia detection: advancements and clinical implications","volume":"25","author":"Abdelrazik","year":"2025","journal-title":"Sensors"},{"key":"B2","doi-asserted-by":"publisher","first-page":"1246746","DOI":"10.3389\/fphys.2023.1246746","article-title":"Deep learning for ECG arrhythmia detection and classification: an overview of progress for period 2017\u20132023","volume":"14","author":"Ansari","year":"2023","journal-title":"Front. Physiol"},{"key":"B3","doi-asserted-by":"publisher","first-page":"334","DOI":"10.1016\/j.jacep.2024.01.008","article-title":"Continuous atrial fibrillation monitoring from photoplethysmography: comparison between supervised deep learning and heuristic signal processing","volume":"10","author":"Antiperovitch","year":"2024","journal-title":"Clin. Electrophysiol"},{"key":"B4","doi-asserted-by":"publisher","first-page":"15054","DOI":"10.1038\/s41598-019-49092-2","article-title":"Atrial fibrillation detection from wrist photoplethysmography signals using smartwatches","volume":"9","author":"Bashar","year":"2019","journal-title":"Sci. Rep"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1342","DOI":"10.1109\/10.959330","article-title":"Do existing measures of poincare plot geometry reflect nonlinear features of heart rate variability?","volume":"48","author":"Brennan","year":"2001","journal-title":"IEEE Trans. Biomed. Eng"},{"key":"B6","doi-asserted-by":"publisher","first-page":"195","DOI":"10.15406\/ijbsbe.2018.04.00125","article-title":"A review on wearable photoplethysmography sensors and their potential future applications in health care","volume":"4","author":"Castaneda","year":"2018","journal-title":"Int. J. Biosens. Bioelectron"},{"key":"B7","doi-asserted-by":"publisher","first-page":"355","DOI":"10.1109\/JPROC.2022.3149785","article-title":"Wearable photoplethysmography for cardiovascular monitoring","volume":"110","author":"Charlton","year":"2022","journal-title":"Proc. IEEE"},{"key":"B8","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/BSN63547.2024.10780734","article-title":"\u201cSmartwatch photoplethysmogram-based atrial fibrillation detection with premature atrial and ventricular contraction differentiation using densely connected convolutional neural networks,\u201d","author":"Chen","year":"2024","journal-title":"2024 IEEE 20th international conference on body sensor networks (BSN)"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1528","DOI":"10.1109\/IEMBS.2007.4352592","article-title":"\u201cA comparative evaluation of adaptive noise cancellation algorithms for minimizing motion artifacts in a forehead-mounted wearable pulse oximeter,\u201d","author":"Comtois","year":"2007","journal-title":"2007 29th annual international conference of the IEEE engineering in medicine and biology society"},{"key":"B10","doi-asserted-by":"publisher","first-page":"e45137","DOI":"10.2196\/45137","article-title":"Accuracy, usability, and adherence of smartwatches for atrial fibrillation detection in older adults after stroke: randomized controlled trial","volume":"7","author":"Ding","year":"2023","journal-title":"JMIR Cardio"},{"key":"B11","doi-asserted-by":"publisher","first-page":"100033","DOI":"10.1016\/j.eswax.2020.100033","article-title":"A review on deep learning methods for ECG arrhythmia classification","volume":"7","author":"Ebrahimi","year":"2020","journal-title":"Expert Syst. Appl. X"},{"key":"B12","doi-asserted-by":"publisher","first-page":"1610","DOI":"10.1109\/JBHI.2019.2950574","article-title":"Detecting atrial fibrillation and atrial flutter in daily life using photoplethysmography data","volume":"24","author":"Eerik\u00e4inen","year":"2019","journal-title":"IEEE J. Biomed. Health Inform"},{"key":"B13","doi-asserted-by":"publisher","first-page":"327","DOI":"10.1016\/j.compbiomed.2018.07.001","article-title":"Automated detection of atrial fibrillation using long short-term memory network with RR interval signals","volume":"102","author":"Faust","year":"2018","journal-title":"Comput. Biol. Med"},{"key":"B14","doi-asserted-by":"publisher","first-page":"861","DOI":"10.1016\/j.patrec.2005.10.010","article-title":"An introduction to ROC analysis","volume":"27","author":"Fawcett","year":"2006","journal-title":"Pattern Recognit. Lett"},{"key":"B15","doi-asserted-by":"publisher","first-page":"687","DOI":"10.1038\/s41746-025-02036-9","article-title":"Integration of artificial intelligence and wearable technology in the management of diabetes and prediabetes","volume":"8","author":"Fraser","year":"2025","journal-title":"NPJ Digit. Med"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1493","DOI":"10.3390\/jcm9051493","article-title":"Population-based screening or targeted screening based on initial clinical risk assessment for atrial fibrillation: a report from the huawei heart study","volume":"9","author":"Guo","year":"2020","journal-title":"J. Clin. Med"},{"key":"B17","doi-asserted-by":"publisher","first-page":"82","DOI":"10.3390\/bios12020082","article-title":"A real-time PPG peak detection method for accurate determination of heart rate during sinus rhythm and cardiac arrhythmia","volume":"12","author":"Han","year":"2022","journal-title":"Biosensors"},{"key":"B18","doi-asserted-by":"publisher","first-page":"5683","DOI":"10.3390\/s20195683","article-title":"Premature atrial and ventricular contraction detection using photoplethysmographic data from a smartwatch","volume":"20","author":"Han","year":"2020","journal-title":"Sensors"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/ICASSP49660.2025.10889502","article-title":"\u201cMulticlass arrhythmia classification using smartwatch photoplethysmography signals collected in real-life settings,\u201d","author":"Han","year":"2025","journal-title":"ICASSP 2025-2025 IEEE international conference on acoustics, speech and signal processing (ICASSP)"},{"key":"B20","doi-asserted-by":"publisher","first-page":"373","DOI":"10.1093\/eurheartj\/ehab648","article-title":"2020 ESC guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS) the task force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) developed with the special contribution of the European Heart Rhythm Association (EHRA) of the esc","volume":"42","author":"Hindricks","year":"2021","journal-title":"Eur. Heart J"},{"key":"B21","doi-asserted-by":"publisher","first-page":"306","DOI":"10.1016\/0013-4694(70)90143-4","article-title":"EEG analysis based on time domain properties","volume":"29","author":"Hjorth","year":"1970","journal-title":"Electroencephalogr. Clin. Neurophysiol"},{"key":"B22","doi-asserted-by":"publisher","first-page":"30164","DOI":"10.1038\/s41598-025-14172-z","article-title":"Digital biomarkers for interstitial glucose prediction in healthy individuals using wearables and machine learning","volume":"15","author":"Huang","year":"2025","journal-title":"Sci. Rep"},{"key":"B23","doi-asserted-by":"publisher","first-page":"3702","DOI":"10.1038\/s41598-022-07783-3","article-title":"Higher heart rate variability as a predictor of atrial fibrillation in patients with hypertension","volume":"12","author":"Kim","year":"2022","journal-title":"Sci. Rep"},{"key":"B24","doi-asserted-by":"publisher","first-page":"1493","DOI":"10.3390\/s20051493","article-title":"Motion artifact reduction in wearable photoplethysmography based on multi-channel sensors with multiple wavelengths","volume":"20","author":"Lee","year":"2020","journal-title":"Sensors"},{"key":"B25","doi-asserted-by":"publisher","first-page":"e023555","DOI":"10.1161\/JAHA.121.023555","article-title":"Multiclass arrhythmia detection and classification from photoplethysmography signals using a deep convolutional neural network","volume":"11","author":"Liu","year":"2022","journal-title":"J. Am. Heart Assoc"},{"key":"B26","doi-asserted-by":"publisher","first-page":"1415","DOI":"10.1161\/CIRCULATIONAHA.122.060291","article-title":"Detection of atrial fibrillation in a large population using wearable devices: the fitbit heart study","volume":"146","author":"Lubitz","year":"2022","journal-title":"Circulation"},{"key":"B27","doi-asserted-by":"publisher","first-page":"4787","DOI":"10.3390\/s24154787","article-title":"Compressed deep learning models for wearable atrial fibrillation detection through attention","volume":"24","author":"Maekynen","year":"2024","journal-title":"Sensors"},{"key":"B28","doi-asserted-by":"publisher","first-page":"232","DOI":"10.1016\/j.jacep.2022.09.011","article-title":"Clinical validation of 5 direct-to-consumer wearable smart devices to detect atrial fibrillation: Basel wearable study","volume":"9","author":"Mannhart","year":"2023","journal-title":"Clin. Electrophysiol"},{"key":"B29","doi-asserted-by":"publisher","first-page":"451","DOI":"10.1038\/s41586-026-10179-2","article-title":"Insulin resistance prediction from wearables and routine blood biomarkers","volume":"652","author":"Metwally","year":"2026","journal-title":"Nature"},{"key":"B30","doi-asserted-by":"publisher","first-page":"378","DOI":"10.1136\/heart.88.4.378","article-title":"The PNNX files: re-examining a widely used heart rate variability measure","volume":"88","author":"Mietus","year":"2002","journal-title":"Heart"},{"key":"B31","doi-asserted-by":"publisher","first-page":"278","DOI":"10.1016\/j.compbiomed.2018.06.002","article-title":"Automated diagnosis of arrhythmia using combination of CNN and LSTM techniques with variable length heart beats","volume":"102","author":"Oh","year":"2018","journal-title":"Comput. Biol. Med"},{"key":"B32","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1038\/s41746-019-0207-9","article-title":"Photoplethysmography based atrial fibrillation detection: a review","volume":"3","author":"Pereira","year":"2020","journal-title":"NPJ Digit. Med"},{"key":"B33","doi-asserted-by":"publisher","first-page":"1909","DOI":"10.1056\/NEJMoa1901183","article-title":"Large-scale assessment of a smartwatch to identify atrial fibrillation","volume":"381","author":"Perez","year":"2019","journal-title":"N. Engl. J. Med"},{"key":"B34","doi-asserted-by":"publisher","first-page":"102194","DOI":"10.1016\/j.bspc.2020.102194","article-title":"Automated atrial fibrillation detection using a hybrid CNN-LSTM network on imbalanced ECG datasets","volume":"63","author":"Petmezas","year":"2021","journal-title":"Biomed. Signal Process. Control"},{"key":"B35","doi-asserted-by":"publisher","first-page":"2982","DOI":"10.1016\/j.jacc.2020.11.010","article-title":"Global burden of cardiovascular diseases and risk factors, 1990-2019: update from the GBD 2019 study","volume":"76","author":"Roth","year":"2020","journal-title":"J. Am. Coll. Cardiol"},{"key":"B36","doi-asserted-by":"publisher","first-page":"618","DOI":"10.1109\/ICCV.2017.74","article-title":"\u201cGrad-CAM: visual explanations from deep networks via gradient-based localization,\u201d","author":"Selvaraju","year":"2017","journal-title":"Proceedings of the IEEE international conference on computer vision"},{"key":"B37","doi-asserted-by":"publisher","first-page":"e45111","DOI":"10.7759\/cureus.45111","article-title":"Evaluation of atrial fibrillation detection in short-term photoplethysmography (PPG) signals using artificial intelligence","volume":"15","author":"Talukdar","year":"2023","journal-title":"Cureus"},{"key":"B38","doi-asserted-by":"publisher","first-page":"109101","DOI":"10.1016\/j.cmpb.2025.109101","article-title":"Lightweight element-wise product enhanced neural network for efficient arrhythmia detection on embedded devices","volume":"273","author":"Tang","year":"2025","journal-title":"Comput. Methods Prog. Biomed"},{"key":"B39","doi-asserted-by":"publisher","first-page":"1043","DOI":"10.1161\/01.CIR.93.5.1043","article-title":"Heart rate variability: standards of measurement, physiological interpretation, and clinical use","volume":"93","year":"1996","journal-title":"Circulation"},{"key":"B40","doi-asserted-by":"publisher","first-page":"116","DOI":"10.1038\/s41746-020-00320-4","article-title":"Multi-task deep learning for cardiac rhythm detection in wearable devices","volume":"3","author":"Torres-Soto","year":"2020","journal-title":"NPJ Digit. Med"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1579","DOI":"10.3390\/s20061579","article-title":"Automatic detection of arrhythmia based on multi-resolution representation of ECG signal","volume":"20","author":"Wang","year":"2020","journal-title":"Sensors"},{"key":"B42","unstructured":"\u201cCardiovascular diseases (CVDs),\u201d\n          \n          World Health Organization Fact Sheet\n          \n          2025"},{"key":"B43","doi-asserted-by":"publisher","first-page":"673","DOI":"10.3390\/s19030673","article-title":"Motion artifact reduction for wrist-worn photoplethysmograph sensors based on different wavelengths","volume":"19","author":"Zhang","year":"2019","journal-title":"Sensors"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1882739\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T05:44:03Z","timestamp":1790574243000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1882739\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,28]]},"references-count":43,"alternative-id":["10.3389\/frai.2026.1882739"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1882739","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,28]]},"article-number":"1882739"}}