Capillary Refill Time (CRT) is a visual method used to evaluate
peripheral perfusion, particularly in low-resource environments.
However, CRT’s repeatability and reproducibility are limited,
particularly for dark-skinned individuals. This paper presents
quantitative CRT measurements with good performance and repeatability
for all Fitzpatrick skin phototypes. Tests were conducted on the forearm
of 22 volunteers, using 7 kPa controlled compression, an RGB video
camera, and white LED cross-polarized light. The CRT was determined by
the time constant of an exponential regression to the green channel mean
pixel intensity. An adaptive algorithm was developed to determine the
best regression region to reduce noise, and incorrect and outlier CRT
readings were flagged from the regression uncertainties. Results showed
80% of measurements were within 20% of the expected CRT value of a
given volunteer, suggesting repeatable and reproducible quantitative CRT
methods could be developed for robust measurements in patient triage,
monitoring, and telehealth.