EEMD. VMD, DMD, and FFT in Remote Photoplethysmography for Contactless Heart Rate and Respiration Rate Measurement

There is a need for fast, accurate, contactless, and continuous vital sign measurements to provide insights into overall health and to enhance medical care in hospitals and other healthcare facilities. Remote Photoplethysmotraphy (rPPG) is a contactless method that can enable quick and non-invasive monitoring of vital signs such as heart rate (HR) and respiration rate (RR). There have been a growing number of studies creating and testing different signal decomposition techniques with rPPG for frequency-based vital measurement of HR and RR; however, there are still numerous shortcomings. Methods including Dynamic Mode Decomposition (DMD) and Variational Mode Decomposition (VMD) are sparsely discussed with respect to rPPG, and have yet to be applied to rPPG for RR estimation. Additionally, there are no comprehensive analyses on these techniques for rPPG, including DMD, VMD, and other techniques such as Ensemble Empirical Mode Decomposition (EEMD) and Fast Fourier Transform (FFT) which are more established in the rPPG space. This study provides a comparative analysis on EEMD, VMD, DMD, and FFT for rPPG HR and RR estimation. DMD outperformed the other methods in HR estimation with an MAE and RMSE of 3.08 and 5.05 beats per minute (bpm), respectively. For RR estimation, VMD performed the best with an MAE and RMSE of 3.36 and 4.08 breaths per minute (bpm), respectively. These results highlight the importance of developing new rPPG methods with DMD and VMD for HR and RR estimation.