Learning of Art Style Using AI and Its Evaluation Based on Psychological Experiments

Abstract

GANs (Generative adversarial networks) is a new AI technology that has the capability of achieving transformation between two image sets. Using GANs we have carried out a comparison between several art sets with different art styles. We have prepared four image sets; a flower image set with Impressionism art style, one with the Western abstract art style, one with Chinese figurative art style, and one with the art style of Naoko Tosa, one of the authors. Using these four sets we have carried out a psychological experiment to evaluate the difference between these four sets. We have found that abstract drawings and figurative drawings are judged to be different, figurative drawings in West and East were judged to be similar, and Naoko Tosa’s artworks are similar to Western abstract artworks.