Heart sound signal recovery based on time series signal prediction using a recurrent neural network in the long short-te
- PDF / 1,514,696 Bytes
- 18 Pages / 439.37 x 666.142 pts Page_size
- 96 Downloads / 184 Views
Heart sound signal recovery based on time series signal prediction using a recurrent neural network in the long short‑term memory model Zhi‑Hao Wang1 · Gwo‑Jiun Horng2 · Tz‑Heng Hsu2 · A. Aripriharta3 · Gwo‑Jia Jong1
© Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract In this research, we propose a method for recovering heart sound signals that comprises the long short-term memory prediction model based on the recurrent neuron network architecture. The complete heart sound signal is used to implement a prediction model to recover damaged or incomplete heart sound signals. Root mean square errors (RMSEs) and Pearson’s correlation coefficients are used for numerical evaluation. The signals of 13 out of 15 subjects are considerably improved, with the RMSE being as low as 0.03 ± 0.04. Using the Pearson correlation coefficients to estimate the degree of signal recovery, the highest coefficient of correlation between the original and recovered time-domain waveforms is 0.93, and that between the corresponding spectra is 0.967. Waveforms and spectra are used to compare the results graphically. The recovered signal more closely fits the original signal than the interfered signal. Additionally, excess frequency components in the recovered spectra are found to be filtered out and important features retained. Thus, the proposed method not only recovers incomplete or disturbed signals but also has the effect of a filter. Keywords Heart sound signal recovery · Long short-term memory (LSTM) · Recurrent neural networks (RNN) · Root mean square error (RMSE)
* Gwo‑Jiun Horng [email protected] 1
Department of Electronic Engineering, National Kaohsiung University of Science and Technology, Kaohsiung, Taiwan
2
Department of Computer Science and Information Engineering, Southern Taiwan University of Science and Technology, Tainan, Taiwan
3
Department of Electrical Engineering, Universitas Negeri Malang, Malang, Indonesia
13
Vol.:(0123456789)
Z.-H. Wang et al.
1 Introduction The heart is one of the most crucial organs in the human body, and heart disease can cause death. The current methods for detecting heart disease include electrocardiography, treadmill electrocardiography, electrophonocardiography, echocardiography, angiocardiography, and single proton computed tomography. Listening to heart sounds is a low-cost method and obtains more information than other methods. Under normal circumstances, heart sounds can be separated into four sounds, namely the first to fourth heart sounds. The first heart sound is the sound of atrioventricular valve closure. An irregular first heart sound may indicate that the patient has atrial fibrillation. A loud or whispering first heart sound indicates that the mitral valve is abnormal. The second heart sound is caused by closure of valves including the aortic valve and pulmonary valve. Usually, the second heart sound comprises two “da” sounds separated by a certain interval. The frequency of heart sounds is affected by the blood flow velocity. The third he
Data Loading...
