聚类分析在彩色像景织物分色处理中的应用

本文在彩色像景织物的图像传统分色基础上,结合数据挖掘中K-means聚类算法的不足之处,提出了一种改进的K-means彩色图像的聚类分析的分色方法和失真度指标,并通过实验对传统分色方法与改进的聚类分色方法的失真度进行比较,表明用该聚类方法具有较小的失真度,能起到与原纹样风格一致的效果。

The Application of Clustering Analysis in Color Separation of the Color Woven Photograph

Based on traditional color separation on images of the color woven photograph,and combined with K-means clustering algorithm, the paper presents an improved K-means color separation method of the clustering analysis and the target of the distortion degree of the color image. By comparison with the two color separation methods, we can found that the method of clustering analysis has little distortion degree. So it can be capable of keeping the same effect according with the original one.