细节增强的水墨扩散模拟

Ink Diffusion Simulation for Detail Enhancement

水墨画是中国传统绘画的一种表现形式,具有悠久的发展历史.为了使得基于图像的水墨画绘制在体现水墨风格的同时,保留更多的必要细节信息,针对现有基于图像的水墨画渲染模型存在的不足,提出一种强化细节的水墨画扩散模拟算法.该算法人工设置水墨物理扩散的时间,以保证显著的水墨风格,同时增加了细节的增强处理;采用先进行图像分割、提取需要的细节信息,然后将细节信息与扩散后的图像进行加权合成,得到最终的效果图.实验结果表明,该算法通过强化关键细节信息使得水墨风格更接近实际创作的效果,且能增强画面的层次感.

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.