In practical electrocardiogram (ECG) monitoring, there are some challenges in reducing the data burden and energy costs. Therefore, compressed sensing (CS) which can conduct under-sampling and reconstruction at the same time is adopted in the ECG monitoring application. Recently, deep learning used in CS methods improves the reconstruction performance significantly and can removes of some of the constraints in traditional CS. In this paper, we propose a deep compressive-sensing scheme for ECG signals, based on modified-Inception block and long short-term memory (LSTM). The framework is comprised of four modules: preprocessing; compression; initial; and final reconstruction. We adaptively compressed the normalized ECG signals, sequentially using three convolutional layers, and reconstructed the signals with a modified Inception block and LSTM. We conducted our experiments on the MIT-BIH Arrhythmia Database and Non-Invasive Fetal ECG Arrhythmia Database to validate the robustness of our model, adopting Signal-to-Noise Ratio (SNR) and percentage Root-mean-square Difference (PRD) as the evaluation metrics. The PRD of our scheme was the lowest and the SNR was the highest at all of the sensing rates in our experiments on both of the databases, and when the sensing rate was higher than 0.5, the PRD was lower than 2%, showing significant improvement in reconstruction performance compared to the comparative methods. Our method also showed good recovering quality in the noisy data.

No abstract

Electrocardiogram (ECG) signals have been widely used to detect cardiac arrhythmia. Visual inspection is not only time consuming, but also may lead to misdiagnosis and affect the prevention or treatment of the disease. Therefore, automatic diagnosis which can greatly improve the diagnosis efficiency and accuracy are needed to assist doctors for arrhythmia diagnosis. Due to the capacity of high resolution，HRNet has attracted extensive attention for classification in recent years. However, HRNet is only designed for two-dimensional image, and thus not suitable for ECG signals classification. In this paper, we propose an arrhythmia classification scheme which is based on the modified HRNet and efficient channel attention (ECA) to classify five arrhythmia types. The proposed scheme first divides the original ECG signal into 5s segments of 1800 sampling points. Then, the segments are input into the improved HRNet network for automatic learning and classification. Extensive simulations have been performed on MIT-BIH database to validate the effectiveness of the proposed scheme. Experimental results have shown that the proposed scheme achieves an average accuracy of 99.86%, which is superior to the benchmarking methods.

No abstract