In the eye of the beholder

Abstract

Most artworks are explicitly created to evoke a strong emotional response. During the centuries there were several art movements which employed different techniques to achieve emotional expressions conveyed by artworks. Yet people were always consistently able to read the emotional messages even from the most abstract paintings. Can a machine learn what makes an artwork emotional? In this work, we consider a set of 500 abstract paintings from Museum of Modern and Contemporary Art of Trento and Rovereto (MART), where each painting was scored as carrying a positive or negative response on a Likert scale of 1-7. We employ a state-of-theart recognition system to learn which statistical patterns are associated with positive and negative emotions. Additionally, we dissect the classification machinery to determine which parts of an image evokes what emotions. This opens new opportunities to research why a specific painting is perceived as emotional. We also demonstrate how quantification of evidence for positive and negative emotions can be used to predict the way in which people observe paintings.

Beauty in abstract paintings: perceptual contrast and statistical properties

“…However, this difference may well be explained by the different test stimuli used in the experiments (homogeneously colored squares vs. artworks). In another recent investigation, Yanulevskaya et al (2012) showed that bright and saturated colors generated positive emotions, while darker colors tended to evoke negative emotions. Moreover, Amirshahi et al (2013) described a strong correlation of beauty ratings with color quantization values and mean color value in a large dataset of figurative paintings.…”

“…Nowadays, image analysis is based on firmly established empirical methods rather than on vague intuitive grounds. For example, computer-assisted algorithms are used to extract image features that characterize aesthetic images (Datta, 2006; Graham and Field, 2007; Redies et al, 2007a, 2012; Amirshahi et al, 2012), to predict emotional responses to paintings (e.g., Yanulevskaya et al, 2012) or to categorize painting styles (Wallraven et al, 2009). It is hoped that, in combination, the two approaches of experimental aesthetics will help us to understand what the specific properties of aesthetic images are and how they elicit brain responses that correlate with aesthetic judgments by the observers (Redies, 2007).…”

“…Forsythe et al (2011) stressed the importance of color as a medium that artists use to communicate with the observer. Yanulevskaya et al (2012) investigated emotional response patterns to certain colors. They revealed a correlation of bright colors and smooth lines with positive emotions, as opposed to dark colors and chaotic texture that go along with negative emotions.…”

The Influence of Art Expertise and Training on Emotion and Preference Ratings for Representational and Abstract Artworks

“…Our group has previously shown that observers with no training in art awarded highly consistent ratings of valence and arousal to abstract works of art [ 39 , 40 ]. In addition, in a computational vision study, it was possible to train a classifier to correctly predict human emotion judgments for abstract artworks based on basic visual features [ 41 ]. These findings are in line with the idea that valence and arousal judgments are formed, at least in part, on the basis of visual features of the artworks (e.g.…”

Large-scale visual sentiment ontology and detectors using adjective noun pairs

“…However, efforts for visual analysis fall far behind. The closest that comes to sentiment analysis for visual content is the analysis of aesthetics [6,15,22], interestingness [12], and affect or emotions [13,21,37,36]. To this end, either low-level features are directly taken to predict emotion [18,13], or indirectly by facial expressions detection [32] or user intent [11].…”