面向多颜色空间的自适应SFCM聚类图像融合分割

Multi-color Space Oriented Adaptive SFCM Clustering Image Fusion Segmentation

论文提出了一种高效的自适应多颜色空间融合方法,在六个不同的颜色空间中选择最佳分割分量,利用直方图和Sugeno模糊C均值聚类算法(SFCM)技术对分割进行初始化,融合了分割结果并合并了相似的区域。利用该算法在Berkeley图像数据库上进行实验,结果表明:与传统的均值漂移(Mean-shift)、FCR、CTM等分割算法相比,该方法分割效果较好且具有一定的实用价值。

An efficient adaptive multi-color space fusion method is proposed in this paper.The optimal segmentation component is selected in six different color spaces,and the segmentation is initialized by histogram and sueno fuzzy c-means clustering(SFCM)technology.The segmentation results are fused and similar regions are merged.Experiments on Berkeley image database show that compared with traditional mean-shift,FCR,CTM and other segmentation algorithms,this method has better segmentation effect and certain practical value.