多变量高斯差分的风格化方法

A stylistic method using extended DOG operator

针对扩展的高斯差分参数可控性的特点,提出了一种基于参数调整的风格化方法.该方法定义输入图像的ETF流,在基于此ETF流的基础上同时进行梯度轴和流线轴的高斯差分滤波.通过对参数进行调整,得到一系列的风格化效果:包括铅笔画、炭笔画、反走样以及反向边缘的效果.实验结果表明,该方法实现简单,风格多样,效果比单一系统更为理想.

In the light of the extended Gauss difference characteristic parameter controllability, this paper put forward a kind of stylistic method based on parameter adjusting. This method defines the input image ETF flow. It meanwhile conducts Gauss difference filter of Gradient axis and the line shaft based on the ETF flow. A series of style effects can obtain through adjusting the parameters, including pencil shading , pastel, negative edges and anti - aliasing. The experimental results show that this method is simple, style diversity. And the effect is better than that of single system.