Research has suggested that observers, when looking at
scenes with multiple illumination levels, tend to avoid
shadowed regions in the images. This result was obtained in
experiments using psychophysics and when using eyetracking
methodology. However, in such demonstrations
participants were asked to estimate color properties. In the
present study, we introduced cognitive tasks applied to
images partially covered with a shadow. Participants, while
looking at the photographed faces, judged age, beauty and
profession of the depicted people. The eye movement
measures (first fixation location, number of fixations, and
dwell time) showed that even under such high-level cognitive
tasks, the visual system has a preference for non-shadow
regions.
