| 摘要: |
| 目的 针对剪纸智能生成中剪纸的地域性和风格性问题，探索一种能够生成特定风格剪纸的智能生成方法。方法 以侗族剪纸为研究对象，对其进行纹样分析，得到其中应用比较广泛的花草纹样。然后，根据花草纹构造特征进行分解，重构出数量较多的保留侗族剪纸风格的图案，同时通过数据增广的方式获得侗族剪纸花草纹数据集。基于该数据集，提出使用生成对抗网络来拟合花草纹数据集的数据分布，生成侗族剪纸风格的纹样。结果 生成的纹样图像质量较高，线条流畅、优美，符合侗族剪纸风格。结论 通过生成对抗网络学习侗族的剪纸风格，取得了较好效果，为特定风格剪纸的智能生成和非物质文化遗产传承提供了新的思路。 |
| 关键词: 剪纸 侗族剪纸 生成对抗网络 智能生成 |
| DOI：10.19554/j.cnki.1001-3563.2021.14.008 |
| 分类号:TB472 |
| 基金项目: |
|
| Intelligent Generation for Specific Style of Paper-Cutting Art |
|
QIAO Jin-hao, XIAO Yi, CUI Yu-dan, JI Tie
|
|
(Hunan University, Changsha 410082, China)
|
| Abstract: |
| Aiming at the challenge of the regionalization and stylization of paper-cutting, this paper explores an intelligent design method which can generate paper-cutting in specific style. Given this, taking Dong minority paper-cutting as the research object, pattern analysis was conducted to obtain the flower and grass patterns which were more widely used among them. Then according to the structural characteristics of flowers patterns, a large number of patterns retaining the Dong paper-cutting style were reconstructed. After that, Data augmented methods were applied to obtain the final Dong minority paper-cutting flower dataset. On this dataset, a generative adversarial network was proposed to fit the data distribution of the flowers pattern, and eventually generate the diversity flower patterns with smooth and graceful lines and Dong paper-cutting style. This paper achieves better results in learning the Dong paper-cutting style through generative adversarial networks, and provides a new idea for the intelligent design of specific style paper-cutting and the inheritance of intangible cultural heritage. |
| Key words: paper-cutting Dong minority paper-cutting generative adversarial network intelligent generation |
渝公网安备 50010702501716号