Noise Rejection for Wearable ECGs Using Modified Frequency Slice Wavelet Transform and Convolutional Neural Networks

Abstract

Progress in wearable techniques makes the long-term daily electrocardiogram (ECG) monitoring possible. However, the long-term wearable ECGs can be significantly contaminated by various noises, which affect the detection and diagnosis of cardiovascular diseases (CVDs). The situation becomes more serious for wearable ECG screening, where the data are huge, and doctors have no way to visually check the signal quality episode-by-episode. Therefore, automatic and accurate noise rejection for the wearable bigdata ECGs is craving. This paper addressed this issue and proposed a noise rejection method for wearable ECGs based on the combination of modified frequency slice wavelet transform (MFSWT) and convolutional neural network (CNN). Wearable ECGs were recorded using the newly developed 12-lead Lenovo smart ECG vest with a sample rate of 500 Hz and a resolution of 16 bit. One thousand 10-s ECG segments were picked up and were manually labeled into three quality types: clinically useful segments with good signal quality (type A), clinically useful segments with poor signal quality (type B), and clinically useless segments (pure noises, type C). Each of the 1,000 10-s ECG segments were transformed into a 2-D timefrequency (T-F) image using the MFSWT, with a pixel size of 200×50. Then, the 2-D grayscale images from MFSWT were fed into a 13-layer CNN model for training the classification models. Results from the standard 5-folder cross-validation showed that the proposed combination method of MFSWT and CNN achieved a highest classification accuracy of 86.3%, which was higher than the comparable methods from continuous wavelet transform (CWT) and artificial neural networks (ANN). The combination of MFSWT and CNN also had a good calculation efficiency. This paper indicated that the combination of MFSWT and CNN is a potential method for automatic identification of noisy segments from wearable ECG recordings. INDEX TERMS Wearable ECG, signal quality assessment (SQA), convolutional neural network (CNN), modified frequency slice wavelet transform (MFSWT).

PVC Recognition for Wearable ECGs Using Modified Frequency Slice Wavelet Transform and Convolutional Neural Network

“…MFSWT can efficiently contain the time-frequency information of ECG in the transformed 2-D images, such as P-wave, QRS complex and T-wave, and were successfully applied in the previous studies [13,14]. A bound signal-adaptive frequency slice function (FSF) was introduced in MFSWT, which can realize the adaptive measurement of signal energy distribution at different observation frequencies.…”

A Deep Learning Approach for Featureless Robust Quality Assessment of Intermittent Atrial Fibrillation Recordings from Portable and Wearable Devices

“…However, when recordings alternating AF and NSR were separately analyzed, its performance was significantly reduced by 15%. Finally, Zhao et al [ 73 ] designed a 2-D CNN with 13 layers to discern among high-quality ECG excerpts, clinically useful ECG segments with poor quality, and clinically useless ECG intervals. After training and testing the algorithm with only 1000 10 s-length ECG segments from healthy subjects, an accuracy about 86% was achieved.…”

Detection of mechanical resonance frequencies for interior permanent magnet synchronous motor servo drives based on wavelet multiresolution filter

“…The supervisory control and data acquisition system based on wavelet multilevel filter bank is developed to extract the singular signals for the voltage variation events in the distributed energy system [5, 6]. The WT‐based algorithms are applied for the feature extraction of electrocardiogram signal to help the disease identification [7, 8]. Moreover, the torque prediction of a three‐phase induction motor is performed with the WT‐based method for the feature analysis of acoustic signals [9].…”