The Cognition of Audience to Artistic Style Transfer

Abstract

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

Article Analyzing the potential influencing factors in the perception of artistic style transfer

“…Instead of transferring all possible feelings of a style, previous models only learn the most dominant visual clues while ignoring the aesthetic experience in the artistic style. If the meaning and emotion of style can be accurately transferred, the output will receive more preference from subjects (Lyu et al 2021). Therefore, apart from the features of the technical level, the representation of semantic and effectiveness levels should also be regarded as the critical ability of artistic style transfer.…”

“…Given the complexity of artistic style, an evaluation framework for style transfer was proposed including universal questions of the degree in color, texture, feeling, and tone (Lyu et al 2021;Wang et al 2021;Sun et al 2022a, b). By comparing the perception results of the main visual elements of painting, it has been proven that there is a strong positive correlation between the perception of color, stroke, and texture and the overall fitness (Lyu et al 2021;Sun et al 2022a, b). Furthermore, audiences will prefer more when features in the semantic and effectiveness layers are accurately transferred (Lyu et al 2021).…”

“…By comparing the perception results of the main visual elements of painting, it has been proven that there is a strong positive correlation between the perception of color, stroke, and texture and the overall fitness (Lyu et al 2021;Sun et al 2022a, b). Furthermore, audiences will prefer more when features in the semantic and effectiveness layers are accurately transferred (Lyu et al 2021). As a result, understanding style from a comprehensive dimension is crucial for enhancing the effect of NST application in art.…”

“…In addition, the aesthetic experience of art is a dynamic process from the external form to internal feeling (Beardsley 1981;Chatterjee 2003;Lin et al 2017). Apart from that (Lyu et al 2021), proved a strong correlation between abstract style cognition and stroke and texture and a similar relationship between realistic style cognition and color. In addition, there are traces concerning the importance of semantic and effectiveness levels (Lyu et al 2020).…”

“…Regarding artistic style transfer, whether high-level features can be transferred is regarded as the critical indicator for evaluating the quality of stylization of output. Based on previous typical samples in (Lyu et al 2021), the study further explored perceptual differences between human artists and AI by calculating the similarity between the optimized sample and the original style from the statistical and subjective scoring.…”