Elevating Pulse Analysis: Variational and Empirical Mode Decomposition Algorithms for Heart Rate Extraction from Plethysmography Waves^*

Abstract

This research emphasizes the use of advanced signal processing methods, including variational and empirical mode decomposition (VMD/EMD) techniques, to extract the instantaneous heart rate from photoplethysmogram (PPG) signals. Plethysmography is a non-invasive method widely used in healthcare for monitoring blood flow dynamics, and accurate heart rate estimation from plethysmography waveforms is crucial for various clinical applications. To improve heart rate extraction accuracy, especially during physical activities, we propose using adaptive decomposition techniques like VMD and EMD. These methods enable us to extract essential features from PPG signals by dividing them into intrinsic mode functions (IMFs), which are vital for heart rate variability (HRV) analysis. We present a comparative analysis of the heart rate extraction outcomes obtained using the annotators from the BUT PPG database, EMD, and VMD approach. Our findings demonstrate a strong correlation between the heart rate extraction results, with a standard deviation of approximately 2.717 between the VMD technique and annotator 2. These results highlight the potential of these advanced signal processing techniques in enhancing heart rate estimation accuracy and contributing to more reliable assessments of cardiovascular health and other physiological aspects.