[论文解读] Complex Deep Learning Models for Denoising of Human Heart ECG Signals

Corneliu Arsene|arXiv (Cornell University)|Aug 27, 2019

ECG Monitoring and Analysis参考文献 24被引用 5

一句话总结

本文提出了一种基于深度学习的去噪框架，用于人体心电图（ECG）信号，采用卷积神经网络（CNN）、长短期记忆网络（LSTM）、受限玻尔兹曼机（RBM）及传统方法。结果表明，通过从真实静息心电图数据生成的合成噪声1秒ECG片段进行训练的CNN，在离线去噪中表现出色，借助预训练模型可有效实现实时应用。

ABSTRACT

Effective and powerful methods for denoising real electrocardiogram (ECG) signals are important for wearable sensors and devices. 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 ECG signals. This paper presents several DL models namely Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), Restricted Boltzmann Machine (RBM) together with the more conventional filtering methods (low pass filtering, high pass filtering, Notch filtering) and the standard wavelet-based technique for denoising EEG signals. These methods are trained, tested and evaluated on different synthetic and real ECG datasets taken from the MIT PhysioNet database and for different simulation conditions (i.e. various lengths of the ECG signals, single or multiple records). The results show the CNN model is a performant model that can be used for off-line denoising ECG applications where it is satisfactory to train on a clean part of an ECG signal from an ECG record, and then to test on the same ECG signal, which would have some high level of noise added to it. However, for real-time applications or near-real time applications, this task becomes more cumbersome, as the clean part of an ECG signal is very probable to be very limited in size. Therefore the solution put forth in this work is to train a CNN model on 1 second ECG noisy artificial multiple heartbeat data (i.e. ECG at effort), which was generated in a first instance based on few sequences of real signal heartbeat ECG data (i.e. ECG at rest). Afterwards it would be possible to use the trained CNN model in real life situations to denoise the ECG signal. This corresponds also to reality, where usually the human is put at rest and the ECG is recorded and then the same human is asked to do some physical exercises and the ECG is recorded at effort. The quality of results is assessed visually but also by using the Root Mean Squared (RMS) and the Signal to Noise Ratio (SNR) measures. All CNN models used an NVIDIA TITAN V Graphical Processing Unit (GPU) with 12 GB RAM, which reduces drastically the computational times. Finally, as an element of novelty, the paper presents also a Design of Experiment (DoE) study which intends to determine the optimal structure of a CNN model, which type of study has not been seen in the literature before.

研究动机与目标

• 开发适用于可穿戴设备与实时监测系统的鲁棒深度学习模型，用于心电信号去噪。
• 通过从真实静息心电图片段生成合成训练数据，解决实时应用中清洁心电图数据有限的挑战。
• 在多种心电图数据集上，对比深度学习模型（CNN、LSTM、RBM）与传统滤波及小波基去噪方法的性能。
• 利用来自MIT PhysioNet数据库的真实与合成心电图数据，评估模型在不同信号长度与噪声条件下的泛化能力。

提出的方法

• 通过在真实静息状态下获取的1秒心电图片段进行增强，生成合成噪声心电图信号，以模拟包含多个人心动悸的信号并添加噪声。
• 在合成的、含噪声的1秒心电图序列上训练卷积神经网络（CNN），以学习去噪表征。
• 在真实高噪声心电图信号上评估训练好的CNN模型，离线设置中使用相同信号进行训练与推理。
• 将CNN的性能与传统方法（低通、高通、陷波滤波）及小波基去噪方法进行对比。
• 以MIT PhysioNet数据库为主要真实心电图数据源，用于跨多个记录与信号长度的训练与评估。
• 通过利用少量清洁心电图数据进行微调，应用迁移学习原理，提升模型在真实世界部署中的性能。

实验结果

研究问题

• RQ1 在离线设置中，基于合成噪声心电图数据训练的CNN模型能否有效对真实高噪声心电图信号进行去噪？
• RQ2 在心电信号去噪中，深度学习模型（CNN、LSTM、RBM）的性能与传统滤波及小波基方法相比如何？
• RQ3 在实时应用中，基于短时清洁心电图片段训练的模型，能在多大程度上泛化至更长的噪声心电图信号？
• RQ4 当可用于训练深度学习模型的清洁心电图数据有限时，数据增强的最佳策略是什么？
• RQ5 从静息心电图片段生成的合成心电图数据能否有效用于训练真实世界去噪任务的模型？

主要发现

• CNN在心电信号去噪中优于传统滤波与小波基方法，尤其在高噪声水平下表现更优。
• 当基于从真实静息心电图数据生成的合成噪声1秒ECG片段进行训练时，CNN在离线去噪中表现出色。
• 该模型展现出良好的泛化能力，经合成数据预训练后，可实现有效实时或近实时去噪。
• 使用合成数据显著降低了对大量清洁心电图数据的依赖，使该方法在可穿戴设备中更具可扩展性。
• 在所测试的深度学习模型中，CNN在不同信号长度与噪声条件下均表现出最一致且优越的性能。
• 所提出的方法可在资源受限的可穿戴系统中实现鲁棒的心电信号去噪，尤其适用于实时处理需求。