首页    期刊浏览 2024年07月06日 星期六
登录注册

文章基本信息

  • 标题:A RELATIVE EVALUATION OF AESTHETIC VALUE FOR CONTEMPORARY ABSTRACT ART CREATED BY COMPUTER CREATIVITY
  • 本地全文:下载
  • 作者:EUN-MI LEE ; MI-KYUNG JOO
  • 期刊名称:Journal of Theoretical and Applied Information Technology
  • 印刷版ISSN:1992-8645
  • 电子版ISSN:1817-3195
  • 出版年度:2017
  • 卷号:95
  • 期号:18
  • 页码:4427
  • 出版社:Journal of Theoretical and Applied
  • 摘要:Research of Computational Creativity is one of the most challenging fields in artificial intelligence of computer systems. Contemporary artists have been created various kinds of visual art based on artificial intelligence such as neural networks, machine learning and so on. However, there is a controversy about whether the artwork created by Computational Creativity has significant aesthetic value against artworks created by humans. In this paper, we evaluate aesthetic value of contemporary abstract artworks created by computer program that Versteeg developed based on neural networks. We relatively analyze the aesthetic value of Computational Creativity against the aesthetic value of human creativity. We conduct experiments with people that 10 pairs of paintings are presented that one is created by Versteegs program and the other is painted by an artist of middle standing. The experimental results show that the relative aesthetic value of artworks created by the computer program is slightly lower than those of artists of middle standing. Also, similar to human artists, we found that Computer Creativity can create masterpiece or failure. As the results created by a human artist fluctuates, Computer Creativity also may create various level of artworks in the view of aesthetic value. Since Computational Creativity research in visual art is in the very early stage, creating artwork by Computational Creativity has potentials to be improved and to surpass humans creativity.
  • 关键词:Aesthetic Value; Abstract Art; Artificial Intelligence; Computer Creativity; Neural Networks
国家哲学社会科学文献中心版权所有