Design of Deep Convolutional Neural Network Architectures for Denoising Electrocardiographic Signals

Abstract

Deep Learning (DL) models have been used extensively in image processing and other domains with great success but only very recently have been used in processing electrocardiogram (ECG) signals. Effective and powerful methods for denoising real ECG signals are important for wearable sensors and devices. This paper presents the design of a DL model, the Convolutional Neural Networks (CNNs), together with the more conventional filtering methods (low pass filtering, high pass filtering, Notch filtering) and the standard wavelet-based technique for denoising ECG signals. This combination shows to produce better results than using only the CNN models. The methods are trained, tested and evaluated on real ECG datasets taken from the MIT PhysioNet database. The results show the CNN model is a performant model that can be used for denoising ECG signals, and also that before applying the CNN model, standard filtering of the original ECG signal has a beneficial effect for the final results. All CNN models used an NVIDIA TITAN V Graphical Processing Unit (GPU) with 12 GB RAM, which reduces drastically the computational times.