基于无参数Markov随机场模型的彩色纹理综合方法

Color Texture Synthesis Based on Nonparametric Markov Random Field

提出一种基于多尺度、无参数Markov随机场模型和KL变换的彩色纹理综合方法。该方法能捕捉原始纹理的高阶统计特性，综合出与原始纹理视觉一致的彩色纹理图像。实验结果证明该方法对彩色纹理图像的综合非常有效。

This paper presents a color texture synthesis method based on multiscale, nonparameter Markov random field model and KL transform. This method is capable of capturing the high-order statistical characteristics of original textures and synthesizing textures visually indistinguishable from original textures. Experiments of color texture synthesis have been done and the results verify that this method can synthesize color textures effectively.