Non-Photorealistic Rendering in Chinese Painting

In this thesis, we present an automatic framework to convert an input image to a stylized painting with Chinese wash painting effect automatically. Some traditional algorithms are improved and some new algorithms are developed as well.

As we known, Chinese ink is a kind of special pigment to render a Chinese painting, and the various gray effects on the painting make sumi-e (Chinese wash painting) distinctive. Additionally, the ink diffusion along the stroke edges is a significant aspect of Chinese wash painting. Moreover, Chinese paper (Xuan paper) is also an important characteristic. In order to simulate the effects described above, we present a rendering method based on an input colour image. At first, Mesn Shift segmentation algorithm is employed to segment and filter the colour image, and after we change the image to greyscale, the similar gray effect of an ink painting is shown. Then, the necessary edges will be detected to wait for linear diffusion. At the same time, the internal part will be diffused by the non-linear filtering. After compositing the edges and body, paper texture is added by image quilting technology at last.

In addition, NPRP (Non-photorealistic rendering from photos) technology is improved in this thesis. The experimental results illustrate that our methods are successful and promising.

More Results:

Figure 1. LEFT: Real Paintings. RIGHT: Simulation Results

Figure 2. LEFT: original image. RIGHT: Simulation result.