Image processing can be used to recognise human emotions based on facial features in static or video images. Some of these involve muscles that are hard or impossible to control voluntarily. In addition pupil size can be used to gauge arousal and in video slight changes in colour due to bloodflow can be used to detect flushing (e.g. embarassment or anger) and heart rate. Typically these physiological metrics need to be used alongside contextual knowledge, as facial expressions are often ambiguous.
Used in Chap. 6: page 83
