Generating Calligraphy Works Based on DiscoGAN
Yuqing Zhao, Guangyuan Fu · IOP Conference Series Materials Science and Engineering · 2019
Abstract Copying the works of a good calligrapher is a necessary step for many beginners. However, due to the large number of Chinese characters, it is a challenge to automatically generate Chinese characters in the style of calligraphers. With the development of deep learning, the recognition of Chinese characters, style migration and other research have achieved remarkable results. However, there are few studies on the generation of calligraphy works, especially on the lack of paired training data. In this work, we propose a method for generating calligraphy works. This method is generally applicable to the generation of calligraphy works of various styles. And the validity of our method is verified by experiments on data sets.