Fig. 7From: COGNIMUSE: a multimodal video database annotated with saliency, events, semantics and emotion with application to summarizationJoint valence-arousal histograms for a intended and b expected emotion (darker signifies higher value)Back to article page