2022
A regularization method to improve adversarial robustness of neural networks for ECG signal classification
Ma L, Liang L. A regularization method to improve adversarial robustness of neural networks for ECG signal classification. Computers In Biology And Medicine 2022, 144: 105345. PMID: 35240379, DOI: 10.1016/j.compbiomed.2022.105345.Peer-Reviewed Original ResearchConceptsDeep neural networksECG signal classificationAdversarial noiseAdversarial attacksNeural networkAdversarial robustness of neural networksSignal classificationAccuracy of deep neural networksRobustness of neural networksField of computer visionImprove DNN robustnessSignal classification applicationsMIT-BIH datasetNovel regularization methodClass label predictionRegularization methodLife-critical applicationsInterpretation of electrocardiograms (ECGAdversarial perturbationsAdversarial robustnessComputer visionClassification applicationsReduction of accuracyPublic datasetsProjected Gradient
2020
Enhance CNN Robustness Against Noises for Classification of 12-Lead ECG with Variable Length
Ma L, Liang L. Enhance CNN Robustness Against Noises for Classification of 12-Lead ECG with Variable Length. 2020, 00: 839-846. DOI: 10.1109/icmla51294.2020.00137.Peer-Reviewed Original ResearchDeep neural networksAdversarial noiseDefense methodsClassification accuracyClassification accuracy of deep neural networksRobustness of deep neural networksAccuracy of deep neural networksImprove robustnessElectrocardiogram signal classificationInterpretation of ECG signalsClass label predictionLife-critical applicationsRobust against noiseClassification taskClassification challengeF1 scoreNeural networkCustom CNNSignal classificationAverage accuracyCNNClean dataECG signalsAutomatic interpretationEvaluation results