Research On The Pigments Diffusing And Stylistic Rendering Of Chinese Painting

The rendering of Chinese painting is always a challenging topic in field of digital art. In order to introduce the traditional art style to colorful digital world and put forward the development of game and animation industry, a deep study is conducted on key problems in interactive rendering and stylistic rendering for Chinese color painting.Firstly, we put forward a pigment model based on KM color optical theory to simulate the pigments using methods in Chinese color painting. With the building of standard pigments palette, the rich color changing art effects of mixture and superposition for multiple pigments are realized. The vegetal pigments and mineral pigments behave very different both in diffusing process and mixing process which is proper in real world.Secondly, we combine the Fick's Law and Einstein-Brown equation to solve the coefficient of diffusion equation. The result shows that radius size of water and pigments particles of one of the most important parameters which affect the diffusion process. The algorithm design makes full use of graphics hardware, so that real time rendering is achieved in the diffusing process of color painting. The huge computing problem in physical modeling method is overcome in this way.Thirdly, we propose a stylistic rendering technique with input of a known artwork. This is very valuable for style reuse of famous ancient artworks. The proposed algorithm is composed of three parts: segmentation of objects in artwork, analysis and extraction of the painting's artistic feature and feature-based 3D Chinese ink-style rendering.By in-depth analysis on different artistic styles among large number of Chinese paintings, the style feature of each objects is conclude into stroke features, color features, and CUN features. The input artwork is segmented by a method which combines the seedless region growing technique and the region merging technique. By improving the criteria of region merging and termination condition, all the relatively independent objects are segmented. Each region of the object is decided the feature category it belongs to among stroke feature, ink color feature or CUN feature. According to this, the specific feature information is extracted to prepare for rendering.Finally, we present a multi-layered component-based rendering mechanism. In the 3D ink painting stylistic rendering process, the rendering method is mainly decided by the feature analysis results. Different components may be applied multiple rendering method. The extracted feature information and surface feature of the 3D model are both used to guide the rendering process. By this way, the artistic style transferring from specific artwork to a 3D scene is accomplished successfully.