论文标题
使用生成对抗网络的端到端中国景观绘画创造
End-to-End Chinese Landscape Painting Creation Using Generative Adversarial Networks
论文作者
论文摘要
当前基于GAN的艺术生成方法由于其依赖条件输入而产生非原始艺术品。在这里，我们提出了素描和绘画gan（Sapgan），这是第一个模型，该模型从端到端生成了中国景观绘画，而没有条件输入。 Sapgan由两个gans组成：用于生成边缘地图的素描组，而油漆gan则是随后的边缘到绘制翻译的涂料。我们的模型在新的中国景观绘画的新数据集上进行了培训，从未用于生成研究。一项242人的视觉图灵测试研究表明，Sapgan绘画被误认为是人类的艺术品，其频率为55％，从基线gan的绘画效果明显优于绘画。我们的作品为真正的机器人艺术生成奠定了基础。
Current GAN-based art generation methods produce unoriginal artwork due to their dependence on conditional input. Here, we propose Sketch-And-Paint GAN (SAPGAN), the first model which generates Chinese landscape paintings from end to end, without conditional input. SAPGAN is composed of two GANs: SketchGAN for generation of edge maps, and PaintGAN for subsequent edge-to-painting translation. Our model is trained on a new dataset of traditional Chinese landscape paintings never before used for generative research. A 242-person Visual Turing Test study reveals that SAPGAN paintings are mistaken as human artwork with 55% frequency, significantly outperforming paintings from baseline GANs. Our work lays a groundwork for truly machine-original art generation.
