Fig. 15From: COGNIMUSE: a multimodal video database annotated with saliency, events, semantics and emotion with application to summarizationSubjective evaluation results by 20 humans. Informativeness (left figure) and enjoyability (right figure) results of AVT summaries at (×5) rate were FUS that denotes the best summary obtained using fusion methods as presented in [12], FF denotes “fast-forward”, while the two newly produced summaries are differentiated by the weight used in the text modalityBack to article page