基于生成式对抗网络的中文字体风格迁移

Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-08-13 Cooperative journals: 《计算机应用研究》

滕少华 孔棱睿

Abstract： A variety of Chinese characters is an important Chinese cultural symbol. Its design and operation is a hard work requiring a lot of professional knowledge. Therefore, for this work, this paper proposed a new method of Chinese font style transfer based on Generative Adversarial Networks. In the experiment, using a generative model based on the residual network structure, performing the adversarial training between the generative model and the discriminative model under the constraint of mean square error. At last, the trained generative model could be used to implement one-to-one and many-to-many style transfer between different Chinese fonts. Experiments show that, compared with the usual L1 regularization method used before, the method proposed in this paper has better performance in font detail generation, simplified Chinese font modeling, improved the fidelity of generated images, and has better flexibility and versatility.

 Hits 1397  Downloads 823  Comment