Deep learning for the classification of atrial fibrillation using wavelet transform-based visual images

被引：1

作者：

Sun, Ling-Chun ^{[1
]}

Lee, Chia-Chiang ^{[2
]}

Ke, Hung-Yen ^{[3
]}

Wei, Chih-Yuan ^{[4
]}

Lin, Ke-Feng ^{[5
,6
]}

Lin, Shih-Sung ^{[7
]}

Hsiu, Hsin ^{[8
]}

Chen, Ping-Nan ^{[9
]}

机构：

[1] Natl Def Med Ctr, Sch Med, Taipei 11490, Taiwan

[2] Natl Taiwan Univ Sci & Technol, Grad Inst Appl Sci & Technol, Taipei 10607, Taiwan

[3] Natl Def Med Ctr, Triserv Gen Hosp, Dept Surg, Div Cardiovasc Surg, Taipei 11490, Taiwan

[4] Natl Def Med Ctr, Grad Inst Life Sci, Taipei 11490, Taiwan

[5] Natl Def Med Ctr, Sch Publ Hlth, Taipei 11490, Taiwan

[6] Natl Def Med Ctr, Triserv Gen Hosp, Med Informat Off, Taipei 11490, Taiwan

[7] Chinese Culture Univ, Dept Comp Sci & Informat Engn, Taipei 11114, Taiwan

[8] Natl Taiwan Univ Sci & Technol, Grad Inst Biomed Engn, Taipei 10607, Taiwan

[9] Natl Def Med Ctr, Dept Biomed Engn, 161,Sec 6,Minquan E Rd, Taipei 11490, Taiwan

来源：

BMC MEDICAL INFORMATICS AND DECISION MAKING | 2025年 / 22卷 / SUPPL5期

关键词：

Atrial fibrillation; MsCWT; Convolutional Neural Network; ResNet101;

D O I：

10.1186/s12911-025-02872-5

中图分类号：

R-058 [];

学科分类号：

摘要：

BackgroundAs the incidence and prevalence of Atrial Fibrillation (AF) proliferate worldwide, the condition has become the epicenter of a plethora of ECG diagnostic research. In recent diagnostic methodologies, Morse Continuous Wavelet Transform (MsCWT) is a feature extraction technique utilized to draw out distinctive attributes of ECG signals. In our study, we explore the employment of MsCWT in the classification of AF with ECG signals in a continuum.ResultsWe present a MsCWT image-based deep learning machine for AF differentiation. For the training, validation, and test sets, we achieved average accuracies of 97.94%, 97.84%, and 91.32%; and overall F1 scores of 97.13%, 96.86%, and 89.41% respectively. Moreover, AUC ROC curves of over 0.99 were obtained for all classes in the training and validation sets; and were over 0.9679 for the test set.ConclusionsTraining deep learning machines for the classification of AF with MsCWT-based images demonstrated to yield favorable outcomes and achieved superior performance amongst studies utilizing the same dataset. Though minimal, the conversion of signals into wavelet form with MsCWT may drastically improve outcomes not only in future ECG signal studies; but all signal-based diagnostics.

引用

页数：11