Convolutional Dendrite Net detects myocardial infarction based on ECG signal measured by flexible sensor

Myocardial infarction (MI) is a kind of heart disease with high mortality, which is caused by long-term myocardial ischemia. To diagnosis MI automatically, an automatic detection method which is based on flexible sensor is proposed in this paper. ECG signal is collected by flexible sensor firstly. After simple preprocessing, ECG is encoded to image by Hilbert curve. Finally, Convolutional Dendrite Net (CDD Net) is used to get the diagnosis results by classifying the image signals. The coding process takes advantage of the characteristics of Hilbert curve. Therefore, the time domain feature of ECG is completely transformed into the shape feature of image. Different from traditional neural networks, CDD pays more attention to the logical combination of features. The method is verified by Physikalisch Technische Bundesanstalt (PTB) dataset. The result shows that the average accuracy of the method is 98.95%. Compared with the existing methods, it has simple structure and higher accuracy. Based on the wearable flexible sensing device, this method can diagnose myocardial infarction anytime and anywhere.