Hemodynamic Estimation Using Sparse Modeling for Facial Thermal Images

Abstract

Facial skin temperature can be measured remotely with an infrared thermography camera to evaluate the physiological and psychological state of a person. To extract specific patterns, a method for analyzing skin temperature distributions that are independent of time series changes is required. Therefore, we applied sparse coding to the facial thermal images. In this study, we attempted to estimate hemodynamics based on the features obtained by applying sparse coding.