基于反向映射的图像间颜色迁移算法仿真

现有的图像间颜色迁移算法普遍存在颜色误迁率高的问题,为此在传统颜色迁移算法的基础上,引入反向映射对颜色迁移算法进行优化。首先通过对图像的特征提取,获得目标图像的主色,对提取结果做平滑处理。通过色彩空间转换和色调反向映射两个步骤,实现图像的形态学变换,得到过渡图像。在过渡图像上做K均值聚类匹配操作,并通过分层迁移和全局迁移两种方式,得到颜色迁移的最优化能量方程。将反向映射颜色迁移算法与传统算法放入仿真环境当中,进行仿真对比实验,通过实验发现传统方法的平均误迁率为12.3%,而反向映射颜色迁移算法的误迁率仅为0.27%。

Simulation of Color Transfer Algorithms Between Images Based on Reverse Mapping

Currently,the color migration algorithm often leads to high false migration rate.Based on the traditional color migration algorithm,the reverse mapping was introduced to optimize the color migration algorithm.Firstly,the main color of target image was obtained by extracting image features.Secondly,the extraction results were smoothed.Through color space conversion and hue reversal mapping,the morphological transformation of image was realized and the transition image was obtained.Thirdly,the transition images were matched by K-means clustering and the optimal energy equation of color migration was obtained by hierarchical migration and global migration.Finally,the color migration algorithm based on reverse mapping was compared with the traditional algorithm.Simulation results show that the average false migration rate of traditional methods is 12.3%,while the false migration rate of color migration algorithm based on reverse mapping is only 0.27%.