Optimum Wavelengths in the Near Infrared for Imaging Photoplethysmography.

Abstract

The purpose of this contribution is to determine the ideal near infrared wavelength bands for monochromatic and dual-band remote heartbeat detection using imaging photoplethysmography (iPPG) of the forehead. Experimental data of 38 healthy volunteers has been recorded and analyzed. For the data acquisition, a fast hyperspectral imager has been used. A new combination approach has been implemented that computes the quotient of the bands and, therefore, reduces motion artifacts. With this dual-band method excellent results (1.67 beats per minute mean deviation from electrocardiogram measurements for 73 recordings) have been obtained using a simple algorithm to analyze images at 799 and 861 nm. It can be concluded that excellent imaging photoplethysmography measurements can be performed at low cost using conventional silicon-based image sensors with invisible light in the near infrared region. This approach is a contribution to the development of non-contact heart rate measurement systems that can be used for medical diagnosis or other applications.