A Deep Multimodal Model for Predicting Affective Responses Evoked by Movies Based on Shot Segmentation

Figure 1

5 s clip versus shot clip. There is little change in the semantic information in the shot clip compared to the 5 s clip.