Interactive GA flock brush for non-photorealistic rendering
SEAL'12 Proceedings of the 9th international conference on Simulated Evolution and Learning
Hi-index | 0.00 |
In non-photorealistic rendering (NPR), sketching is very important to express the preliminary state of an idea. Most current NPR techniques to generate sketching style perform rendering using strokes with different size and shape. In this paper a non-stroke based method is proposed which produces sketching style of painting from source photos. This method preserves only the essential characteristics of the source photo, the result is expressive. To achieve this, we develop a feature preserved filter which preserves the important features of an image. The filter can also remove the illumination effect contained in source photos. The histogram of the filtered image can be approximated well by the normal distribution. Then we design a transformation function to generate a final binary image in which only feature information is left. Namely, the irrelevant information is removed. We also present results of some experiments with this method.