Emonotate: Development a Tool for Collecting Appraisal Trajectories for Time-Series Media Contents

Abstract

In the field of storytelling and filmmaking, we are often aware of the viewer's subjective state, such as emotional state and so on. A similar concept is direction, which is often accumulated by the heuristic findings of creators. The issue is that the emotional state of the viewer is subjective, making it hard for computational approach. Our approach is utilizing appraisal trajectory. The appraisal trajectory is a line chart drawn the viewer's own emotional state from the beginning to the end of the content. We think that the appraisal trajectory enables us to handle the viewer's emotional state with a computer. We think to need accumulate basic findings about subjective trajectories; therefore, we develop a tool to collect subjective trajectories: emonotate. This tool provides some features: input and store the various emotional states of viewers and viewers intuitively agreeable input by visual feedback of arbitrary appraisal trajectories. In order to evaluate the validity, we conduct the experiment of 104 participants drawing subjective trajectories based on protagonist emotional state about 2 animation contents and the subjective quantity is the protagonist emotional state. Both works yielded intuitively correct results, demonstrating the realism of our tool. We think our experimental facts indicate the possibility of our tool collecting subjective trajectories and the issues.