Research Publications

Pupillary response/pupil dilation has been widely accepted as a physiological index of cognitive workload, which can be reliably measured by video eye tracker. It can provide highly relevant cognitive load information. However, in practice pupillary response can be affected by various factors unrelated to workload, such as luminance change and emotional arousal, which might make pupil dilation fail to measure cognitive workload in video based human computer interaction. Dynamic feature extraction from pupillary response data is investigated in this work to robustly index cognitive workload under varying luminance and emotion conditions.