A. Zlatintsi, P. Koutras, G. Evangelopoulos, N. Malandrakis, N. Efthymiou, K. Pastra, A. Potamianos, and P. Maragos, " COGNIMUSE: a multimodal video database annotated with saliency, events, semantics and emotion with application to summarization", EURASIP Journal on Image and Video Processing, vol. 2017, p. 54, Jan. 2017.
In the figure above, the real-time Emotion of users/expert is exposed for each Movie as measured in the database.
The "experienced" emotion ratings are the ones labeled by users and represent their individual subjective experience. There are 7 different users and most of them have annotated each movie twice. The names follow the subject_iteration convention forming the following individual streams of emotion: subj1_1, subj1_2, subj2_1, subj2_2, subj3_1, subj4_1, subj4_2, subj5_1, subj5_2, subj6_1, subj7_1 and subj7_2. Each stream contains the Valence and Arousal measured continuously (every 40 msec - 1 video frame).
Furthermore, the "intended" emotion annotations are labeled by an expert and represent more of the director's intent. It was done 3 times creating the additional streams denoted as: intended_1, intended_2 and intended_3.
The total emotion streams (having duration approximately 30 minutes) of the users/expert for the selected Movie can be seen in the following interactive figures.