| CPC A61B 5/352 (2021.01) [A61B 5/1102 (2013.01); A61B 5/308 (2021.01)] | 19 Claims |
|
1. An electrocardiogram (ECG) monitoring device comprising:
a processor; and
a memory coupled to the processor,
wherein the processor performs a method comprising:
acquiring a vibration signal, including a heart vibration of a person to be observed, by detecting a vibration transmitted through an instrument in a non-contact or non-invasive method using at least one vibration meter sensor attached to the instrument at which the person to be observed is positioned;
extracting a seismocardiography signal (SCG signal) generated by receiving the vibration signal and filtering a predetermined frequency band from the received vibration signal; and
estimating a pattern of the SCG signal and generating an electrocardiogram signal (ECG signal) of a pattern corresponding to the estimated pattern of the SCG signal using a bidirectional long-short term memory (Bi-LSTM) neural network,
wherein the extracting the SCG signal further includes:
performing low pass filtering in order to remove a frequency band exceeding a predetermined first frequency by receiving the vibration signal; and
performing high pass filtering in order to remove a frequency band lower than a predetermined second frequency by receiving the low-pass-filtered signal
wherein the Bi-LSTM neural network is learned in advance using a learning SCG signal and a learning ECG signal synchronized so that a peak of the learning SCG signal and an R peak of the learning ECG signal corresponding thereto in learning data including a plurality of learning SCG signals and a plurality of learning ECG signals corresponding thereto acquired in advance appear at the same time point.
|