摘要
针对非真实感绘制不仅要能有效地表达图像中目标的形状和颜色信息,还必须处理好图像中细节的变化,提出了自适应画笔分布的非真实感绘制方法。通过使用源图像亮度图像的边缘强度来指导画笔的放置,使得画笔分布能够随着源图像中细节的丰富程度而自适应变化。算法运用源图像的梯度向量场控制画笔的运动,使得画笔具有形状、颜色信息的同时,具备了一定的方向信息。同时算法采用“误差扩散”技术对粗画笔的绘制结果进行瑕疵平滑,使得绘制效果更加自然真实。
A good non-photorealistic rendering （NPR） must be able to not only convey the objects＇ shapes and color effectively, but also deal well with the change of details. A NPR method based on adaptive distributions of strokes is proposed. This method can effectively express different level of details in the surface of images and make strokes distributions change with the surface details by directing the strokes＇ placements with the edge signal energy of the source image＇s luminance image. Using source image＇s gradient field can control strokes＇ movement. That can provide an effective aid to express the objects＇ shapes. Meanwhile, ＂error diffusion＂ is used to smooth image flaws brought by rendering with large strokes, making more natural rendering results.
出处
《工程图学学报》
CSCD
北大核心
2006年第2期58-63,共6页
Journal of Engineering Graphics
基金
高等学校博士点学科专项基金资助项目(20020288024)
南京理工大学青年学者基金资助项目(Njust200401)
关键词
计算机应用
非真实感绘制
边缘强度
误差扩散
自适应
computer application
non-photorealistic rendering
edge energy
error diffusion
adaptive
