IEEE Access · 2019 · 101 citations · 51 references
ComputingBiomedical EngineeringCardiologyBiomedical Signal ProcessingComputer VisionFeature EngineeringCardiac Signal ProcessingElectrocardiogram AnalysisElectrocardiographyMedical Image AnalysisPrincipal Component AnalysisSignal ProcessingAutomated Ecg InterpretationCardiac Feature ExtractionComputer-aided Electrocardiogram InterpretationEcg Feature ExtractionElectrocardiogram Feature ExtractionMyocardial Infarction DetectionMyocardial Infarction ImagingNoise Removal
Myocardial infarction (MI) is a deadly disease that threatens human life worldwide, and it is essential to save threatened lives with early detection of MI. The electrocardiogram (ECG), which records the electrical activity presented in the heart, is used for the prevention and treatment of heart disease such as MI. However, it remains a challenge to visually interpret the ECG signals because of their small amplitude and duration. Inspired by the development in computer vision, we try to explore a novel approach for automatic detection of MI by imaging ECG signals without noise removal. In this paper, the ECG time series is first transformed into images using the Gramian Angular Difference Field (GADF) method. Subsequently, the processed images are subjected to the principal component analysis network (PCANet) to extract sparse high-dimensional features, which are easy to perform well in linear classifiers. We carried out several sets of experiments to test the effectiveness of our algorithm. The overall accuracy of 99.49%, the sensitivity of 99.78%, and the specificity of 98.08% are achieved in class-oriented experiments using original ECG beats. The accuracy even rises over 1% compared with the denoising one; Moreover, we also achieved favorable performance for the patient-specific experiment (accuracy of 93.17%, sensitivity of 93.91%, and specificity of 89.20%). The results of the experiments indicate that our model is an effective way to detect MI using raw ECG signals.
51
Corinna Cortes, Vladimir Vapnik · Machine Learning · 1995 · 40.5K citations · Full text
Corinna Cortes, Vladimir Vapnik · Machine Learning · 1995 · 33K citations · Full text
Learning representations by back-propagating errors
David E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams · Nature · 1986 · 30.9K citations
A Fast Learning Algorithm for Deep Belief Nets
Geoffrey E. Hinton, Simon Osindero, Yee‐Whye Teh · Neural Computation · 2006 · 16.4K citations
PhysioBank, PhysioToolkit, and PhysioNet
Ary L. Goldberger, Luı́s A. Nunes Amaral, Leon Glass et al. · Circulation · 2000 · 14.5K citations · Full text