绘画艺术图像的计算美学: 研究前沿与展望

绘画艺术是人类艺术创作的重要组成部分, 绘画艺术图像的计算美学是利用机器实现可计算的人类审美过程, 其在大规模绘画的自动化分析和机器对感性的计算建模上具有重要的应用价值和科学意义. 针对其交叉学科的特点, 本文首次从人类审美的感知、认知和评价三个关键过程出发, 将绘画艺术图像的计算美学研究完整地归纳为属性识别、内容理解和美学评价三方面研究内容, 对其中的问题建模、数据获取和前沿方法等关键科学问题进行了归纳总结, 并对绘画计算美学的三方面研究内容进行了对比、思考和展望.

Computational Aesthetics of Fine Art Paintings: The State of the Art and Outlook

Fine art painting is an essential component of art. The computational aesthetics of fine art painting is a computable human aesthetic process realized by machines, which has significant application value and scientific significance in the automatic analysis of large-scale paintings and computational modeling for aesthetic. Given its interdisciplinary characteristics, for the first time, the computational aesthetics of fine art paintings is completely summarized into three aspects: Attribute recognition, content understanding, and aesthetic judgments according to the key processes of human aesthetics that include perception, cognition, and evaluation. The key scientific issues involved in each aspect are summarized, such as problem modeling, data acquisition, and frontier methods. Also, the three research contents of computational aesthetics of fine art painting are compared, and the future development of this field is discussed.