模糊颜色聚类在颜色传输中的应用  

Application of Fuzzy Color Cluster in Color Transfer between Images

作　　者：钱小燕[1] 肖亮[1] 吴慧中[1] 

机构地区：[1]南京理工大学计算机科学与技术系,南京210094

出　　处：《计算机辅助设计与图形学学报》2006年第9期1332-1336,共5页Journal of Computer-Aided Design & Computer Graphics

基　　金：高等学校博士点学科专项基金(20020288024);南京理工大学青年学者基金(Njust200401);长三角重大科技联合攻关项目(BE2004400)

摘　　要：提出一种自动的、基于模糊颜色聚类的自适应颜色传输算法.首先将源图像和目标图像从RGB空间转换到ιαβ空间,在ιαβ空间对2幅图像进行模糊颜色聚类分析;根据目标图像特征属性对每个聚类域计算相应的匹配权值,图像间匹配权值相近的聚类域确定为图像传输的匹配方向;最后将隶属度控制引入传输方程,实现图像间自适应颜色传输.实验结果表明,该算法比原有颜色传输算法的传输效果更加平滑、自然.An automatic color transfer algorithm based on fuzzy color clustering is proposed in this paper. By the algorithm, the source and target images are first transformed into tαβ space from RGB space. Then fuzzy color clustering is applied to obtain different clustering fields. According to the target statistical properties, the matching weights are calculated to determine the direction of local color transfer. Finally, membership degree control is introduced to color transfer equations to realize the adaptive color transfer between images. The experimental results show that more natural and smoother effects could be achieved by the algorithm proposed in the paper than the classical color transfer algorithms.

关 键 词：颜色传输 聚类分析 颜色空间 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]