Ink Diffusion Simulation for Detail Enhancement

Ink painting is a typical Chinese traditional painting method with a long history.In order to present ink styles and necessary details in image-based ink paintings, and avoid the defects of current image-based ink painting rendering models, we propose an ink simulation method based on image detail enhancement.This method ensures salient diffused ink styles by manually setting the diffusion time of the physically-based ink diffusion model.Moreover, a detail enhancement process is added to retain more details.In our method, image segmentation is performed at first, and then the necessary details are extracted.Finally, the weighted synthesis of the details and the diffusion image is performed. Experiments show that by enhancing the necessary image details and hierarchical perception, we can convert a color image into an ink painting with more realistic diffusion ink styles.