EmoTune - Changing Emotional Response to Images

Tuning an arbitrary input image (center) to evoke a rather negative (left) or positive (right) emotional response. Based on a single input parameter the EmoTune filter alters the brightness, saturation and color temperature to steadily increase the valence of an observer.

Katharina Schwarz, Christian Fuchs, Manuel Finckh and Hendrik P. A. Lensch

University of Tübingen

CIC Color and Imaging Conference 2017 (oral presentation) - Best Student Paper Award.

Images and videos can be touching, triggering emotional responses or even changing the mood of the observer. The response is influenced both by the image content as well as the appearance of the image. In this paper, we investigate how solely simple image processing, i.e., modifying the brightness, the saturation, or the color temperature, actually changes the emotional perception. We collect empirical data on images associated with emotion labels and analyze the valence ratings of the different modifications and their strengths. We show that these relationships tend to be linear only in a limited range while sometimes stronger modifications lead to the opposite effect. Pushing the modifiers towards the boundaries we derive from those ranges and combining them successfully shifted the emotional affect on 92% on around 80 samples. From these findings we derive our EmoTune filter which allows for almost linear control by combining specific modes and demonstrate successful application to both images and videos.

Paper

Bibtex

 @inproceeding{schwarz:CIC2017:EmoTune, 
  author    =  {Katharina Schwarz and 
                       Christian Fuchs and 
                       Manuel Finckh and
                       Hendrik P. A. Lensch 
                      }, 
  title     =      {{EmoTune - Changing Emotional Response to Images}}, 
  booktitle = {Color and Imaging Conference (CIC)},
  year      =    {2017}
}