Infant heart rate estimation based on non-contact UV photoplethysmography

Abstract

Non-contact cardio signal monitoring is a requisite for premature infants, as adhesive sensors and electrodes such as that for an electrocardiogram can damage the epidermis. Many types of research on infants in the newborn intensive care unit (NICU) are still in their early stages with many challenges. Therefore, this study aims to use a digital camera as a non-contact photoplethysmography imaging method to measure and compare the heart rate from two experiments (without/with ultraviolet rays (UV)) inside the intensive care unit. The two experiments (without/with UV) rays yielded promising results in comparison with the reference measurements obtained from ten infants. The results reveal a robust correlation using the pearson correlation coefficients (PCC) of 0.99 (for without UV) and 0.96 (for with UV). Also, the spearman correlation coefficient (SCC) test of 0.97 and 0.95 for two experiments, respectively Bland–Altman analysis revealed a strong relationship between measured and reference readings from without and with UV. The experimental results obtained a low error rate for mean absolute error (MAE), and root mean squared error (RMSE). Therefore, this work presents a new aspect of non-contact monitoring, showing promising performance for upcoming medical applications.