In this work, we investigate
the methods for analysis of information from facial thermal video for
non-intrusive detecting stress and anxiety.
We designed tasks to induce mental stress and anxiety, and built a
database of facial thermal video for 25 subjects. Using the data and the facial
landmarks information, we extracted valuable physiological features from
thermal videos and estimate the subjects’ mental states. Our experimental
results show that stress and anxiety can be efficiently detected from
non-intrusive thermal videos.