The Cognition of Audience to Artistic Style Transfer

Artificial Intelligence (AI) is becoming more popular in various fields, including the area of art creation. Advances AI technology bring new opportunities and challenges creation, experience, appreciation art. The neural style transfer (NST) realizes intelligent conversion any artistic using networks. However, product cognition that involving from visual to feel. purpose this paper study factors affecting audience cognitive difference preference on transfer. Those are discussed investigate application generator model Therefore, based artist’s encoding attributes (color, stroke, texture) audience’s decoding levels (technical, semantic, effectiveness), proposed a framework evaluate perspective cognition. Thirty-one subjects with background art, aesthetics, design were recruited participate experiment. experimental process consists four groups, Fauvism, Expressionism, Cubism, Renaissance. According finding study, participants can still recognize different styles after transferred by Besides, features texture stroke impact perception fitness than color. may prefer samples high semantic effectiveness levels. above indicates through automated routine work, be kept transferred.