Noise reduction in magnetocardiograph based on time-shift PCA just using measurement data

Background Magnetocardiograms (MCGs) have become increasingly relevant for clinical research, due to its potential to detect early stages of heart disease. However, it is difficult to assess heart activity precisely without some form of noise reduction, because MCG measurements are extremely small compared to environmental magnetic noise. One of solutions that can suppress the noise is the use of a digital signal processing (DSP) method. The finite impulse response (FIR) filter is a well-known method in reducing noise via DSP. However, FIR filters have various issues such as distorted waveforms, generation of phase differences, and reduced signal peaks. Hence, a noise reduction method using a time-shift principal component analysis (PCA) [1], [2], [3] is considered. This method reduces noise by subtracting reconstructed noise by reference data from measurement data. This method has to need reference sensor, so this method cannot use without reference sensor system. We propose a new time-shift PCA method without reference sensor in order to apply the time-shift PCA without reference sensor systems.