End-to-End Chinese Landscape Painting Creation Using Generative Adversarial Networks
This work addresses the need for machine-original art generation, particularly for Chinese landscape painting creation, though it is incremental as it builds on existing GAN methods.
The paper tackles the problem of unoriginal artwork in GAN-based art generation by proposing SAPGAN, the first end-to-end model that generates Chinese landscape paintings without conditional input, achieving 55% human-mistaken frequency in a Visual Turing Test.
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.