计算机辅助绘制工笔人物画效果
Computer-Aided Rendering of Chinese Meticulous Figure Paintings
-
摘要: 工笔画是我国传统绘画艺术,画面工整细致,但物体的渲染费工费时.为此提出一种计算机辅助绘制算法来加速物体渲染并保持工笔画的渲染视觉特征.该算法对工笔人物画的线描画稿图像提取轮廓,通过交互得到绘画人物的脸部、头发、四肢、衣服所在区域.对于脸部,先交互指定简单的辅助线定义明暗特征位置,再根据一个阻尼弹簧扩散模型进行扩散,生成整个区域的明暗变化灰度图;对于四肢和衣服,提出一种自动确定其明暗关系的算法;对于头发,将线描画稿图像的线条细节作为头发的高频分量,然后对高频分量进行高斯模糊,得到头发的低频分量,再将两者融合得到最终的头发渲染效果;对于其他如头饰等小的部位,采用简单交互实现绘制.将采用文中方法得到的渲染结果与人工渲染作品相比较,效果令人满意.Abstract: Meticulous figure painting is a traditional Chinese art form in which objects are depicted with fine details. Coloring objects in meticulous figure painting by hand is however tedious and time consuming. This paper presents a rendering algorithm to speed up this coloring process and the resulting renders still retain the style of Chinese meticulous figure paintings. We first extract the contour line and obtain regions for the face, hair, limbs and clothes of painted figures interactively. For the face region, we add simple auxiliary lines interactively to specify the light and dark areas, and then use a simple iterative dampened-spring diffuser to generate the gray scale shading of the whole area. For the regions of limbs and clothes, an automatic method is proposed to render their shadings. For the hair region, we take the hand drawn hair lines as the high frequency components, and then smooth it with a Gaussian filter to obtain the low frequency components, and blend them to get the final rendered hair. For other small parts such as headwear, we process them with simple interactions. Compared with hand painted art works, our rendering results look satisfactory.