目前水墨画渲染的算法都没有考虑立体视觉的实现，并且相关算法都过多的依赖于人交互。本项目提出能满足立体视觉要求的三维水墨画自动渲染算法研究。将水墨画渲染过程分为：笔触提取、水墨扩散效果合成，水墨画纹理生成，纹理映射等几个过程。实现水墨画笔触的帧间连续性和水墨画的立体显示是本项目要解决的核心内容。为了产生双眼深度视差，研究利用纹理映射的方法直接在模型网格上生成水墨扩散效果。为保持渲染结果的帧间（左右眼）一致性，采用图像合成的方法模拟水墨扩散效果；将水墨扩散区域分为基本扩散区和不规则边界区，分别合成这两个区域。本项目的研究成果不仅可以实现三维水墨的立体显示，而且能为游戏渲染引擎和数字美术平台的开发提供技术支撑。

The current 3D rendering algorithms for Chinese-painting-style paintings are not considered stereovision applications, and the need too much interactive operations. The project will investigate the automatic rendering methods that meet the applications of 3D stereovision for Chinese wash-ink paintings. The rendering process is divided into four steps: brush strokes segmentation, ink diffusion synthesis, ink-style textures synthesis and texture mapping. Maintaining frame-to-frame coherence of wash-ink strokes and realizing stereovision display are the key problems of this research. To generate the perspective parallax of double eyes when rendering 3D Chinese painting, we directly generate ink diffusion on 3D meshes using texture mapping. To maitain temperal coherence of wash ink strokes, we present a novel image-based method to synthesize ink diffusion effects. Ink diffusion area is divided into two parts: Primary region and irregular border region. The two parts are synthesized respectively. This research can not only realize stereoscopic display of three-dimensional Chinese painting,but also provide technical support for the developments of game rendering engine and digital art platforms.