Objective To design a deep learning model that balances model complexity and performance to enable its integration into wearable ECG monitoring devices for automated diagnosis of atrial fibrillation. Methods This study was performed based on data from 84 patients with atrial fibrillation, 25 patients with atrial fibrillation, and 18 subjects without obvious arrhythmia collected from the publicly available datasets LTAFDB, AFDB, and NSRDB, respectively. A lightweight attention network based on depthwise separable convolution and fusion of channel-spatial information, namely DSC-AttNet, was proposed. Depthwise separable convolution was introduced to replace standard convolution and reduce model parameters and computational complexity to realize high efficiency and light weight of the model. The multilayer hybrid attention mechanism was embedded to compute the attentional weights of the channels and spatial information at different scales to improve the feature expression ability of the model. Ten-fold cross-validation was performed on LTAFDB, and external independent testing was conducted on AFDB and NSRDB datasets. Results DSC-AttNet achieved a ten-fold average accuracy of 97.33% and a precision of 97.30% on the test set, both of which outperformed the other 4 comparison models as well as the 3 classical models. The accuracy of the model on the external test set reached 92.78%, better than those of the 3 classical models. The number of parameters of DSC-AttNet was 1.01M, and the computational volume was 27.19G, both smaller than the 3 classical models. Conclusion This proposed method has a smaller complexity, achieves better classification performance, and has a better generalization ability for atrial fibrillation classification.
ShiSB, TangYH, ZhaoQY, et al. Prevalence and risk of atrial fibrillation in China: a national cross-sectional epidemiological study[J]. Lancet Reg Health West Pac, 2022, 23: 100439.
HalcoxJPJ, WarehamK, CardewA, et al. Assessment of remote heart rhythm sampling using the AliveCor heart monitor to screen for atrial fibrillation: the REHEARSE-AF study[J]. Circulation, 2017, 136(19): 1784-94.
[4]
DaqrouqK, AlkhateebA, AjourMN, et al. Neural network and wavelet average framing percentage energy for atrial fibrillation classification[J]. Comput Methods Programs Biomed, 2014, 113(3): 919-26.
[5]
FuadahYN, LimKM. Optimal classification of atrial fibrillation and congestive heart failure using machine learning[J]. Front Physiol, 2022, 12: 761013.
[6]
BasharSK, HossainMB, DingE, et al. Atrial fibrillation detection during sepsis: study on mimic iii icu data[J]. IEEE J Biomed Health Inform, 2020, 24(11): 3124-35.
[7]
BasharSK, HanD, ZieneddinF, et al. Novel density poincaré plot based machine learning method to detect atrial fibrillation from premature atrial/ventricular contractions[J]. IEEE Trans Biomed Eng, 2021, 68(2): 448-60.
[8]
García-IslaG, MainardiL, CorinoVDA. A detector for premature atrial and ventricular complexes[J]. Front Physiol, 2021, 12: 678558.
[9]
GuptaK, BajajV, Ahmad AnsariI. Atrial fibrillation detection using electrocardiogram signal input to LMD and ensemble classifier[J]. IEEE Sens Lett, 2023, 7(6): 7002904.
[10]
AsgariS, MehrniaA, MoussaviM. Automatic detection of atrial fibrillation using stationary wavelet transform and support vector machine[J]. Comput Biol Med, 2015, 60: 132-42.
[11]
Abdul-KadirNA, Mat SafriN, OthmanMA. Dynamic ECG features for atrial fibrillation recognition[J]. Comput Methods Programs Biomed, 2016, 136: 143-50.
[12]
LuongoG, ReesF, NairnD, et al. Machine learning using a single-lead ECG to identify patients with atrial fibrillation-induced heart failure[J]. Front Cardiovasc Med, 2022, 9: 812719.
[13]
KalidasV, TamilLS. Detection of atrial fibrillation using discrete-state Markov models and Random Forests[J]. Comput Biol Med, 2019, 113: 103386.
[14]
LiXY, ShiXL, HandaBS, et al. Classification of fibrillation organisation using electrocardiograms to guide mechanism-directed treatments[J]. Front Physiol, 2021, 12: 712454.
[15]
MateoJ, Joaquín RietaJ. Radial basis function neural networks applied to efficient QRST cancellation in atrial fibrillation[J]. Comput Biol Med, 2013, 43(2): 154-63.
[16]
ChenB, MasloveDM, CurranJD, et al. A deep learning model for the classification of atrial fibrillation in critically ill patients[J]. Intensive Care Med Exp, 2023, 11(1): 2.
[17]
PourbabaeeB, RoshtkhariMJ, KhorasaniK. Deep convolutional neural networks and learning ECG features for screening paroxysmal atrial fibrillation patients[J]. IEEE Trans Syst Man Cybern Syst, 2017, 48(12): 2095-104.
[18]
FaustO, ShenfieldA, KareemM, et al. Automated detection of atrial fibrillation using long short-term memory network with RR interval signals[J]. Comput Biol Med, 2018, 102: 327-35.
[19]
KumarD, PeimankarA, SharmaK, et al. Deepaware: a hybrid deep learning and context-aware heuristics-based model for atrial fibrillation detection[J]. Comput Methods Programs Biomed, 2022, 221: 106899.
[20]
LaghariAA, SunYQ, AlhusseinM, et al. Deep residual-dense network based on bidirectional recurrent neural network for atrial fibrillation detection[J]. Sci Rep, 2023, 13(1): 15109.
[21]
IandolaFN, HanS, MoskewiczMW, et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size[EB/OL]. 2016: 1602.07360.
[22]
HowardAG, ZhuML, ChenB, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL]. 2017: 1704.04861.
[23]
LiYJ, ChenM, WangY, et al. Diagnosis of atrial fibrillation based on lightweight detail-semantic network[J]. Biomed Signal Process Contr, 2023, 85: 105025.
[24]
PetrutiuS, SahakianAV, SwirynS. Abrupt changes in fibrillatory wave characteristics at the termination of paroxysmal atrial fibrillation in humans[J]. Europace, 2007, 9(7): 466-70.
[25]
MoodyGB, MarkRR. New method for detecting atrial fibrillation using r-r intervals[J]. Comput Cardiol, 1983: 227-30.
[26]
GoldbergerAL, AmaralLA, GlassL, et al. PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals[J]. Circulation, 2000, 101(23): E215-20.
HamiltonP. Open source ECG analysis[C]//Computers in Cardiology. September 22-25, 2002, Memphis, TN, USA. IEEE, 2002: 101-4.
[29]
ChristovII. Real time electrocardiogram QRS detection using combined adaptive threshold[J]. Biomed Eng Online, 2004, 3(1): 28.
[30]
WooS, ParkJ, LeeJY, et al. CBAM: convolutional block attention module[C]//Computer Vision-ECCV 2018. Cham: Springer International Publishing, 2018: 3-19.
[31]
LiuS, WangAG, DengXT, et al. MGNN: a multiscale grouped convolutional neural network for efficient atrial fibrillation detection[J]. Comput Biol Med, 2022, 148: 105863.
[32]
PrabhakararaoE, DandapatS. Atrial fibrillation burden estimation using multi-task deep convolutional neural network[J]. IEEE J Biomed Health Inform, 2022, 26(12): 5992-6002.
[33]
SunYH, LiYJ, ChenM, et al. IMC-ResNet: Atrial fibrillation detection method based on interlayer multiscale coupling[J]. Biomed Signal Process Contr, 2024, 97: 106683.
[34]
MousaviSS, AfghahF, RaziA, et al. ECGNET: Learning where to attend for detection of atrial fibrillation with deep visual attention[J]. IEEE EMBS Int Conf Biomed Health Inform, 2019: 10.1109/BHI.2019.8834637.
[35]
CholletF. Xception: deep learning with depthwise separable convolutions[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). July 21-26, 2017, Honolulu, HI, USA. IEEE, 2017: 1800-7.
[36]
ChenLC, ZhuYK, PapandreouG, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]// Computer Vision-ECCV 2018. Cham: Springer International Publishing, 2018: 833-51.