基于改进C均值聚类的自适应图像分色算法

图像分色在纺织和印刷等行业中有着广泛而重要的应用，其目的是用尽量少的颜色来描述一幅真彩色图像，使得到的图像与原图像尽可能的接近。该文提出了一种基于改进C均值聚类的自适应图像分色算法。该算法首先随机产生一张颜色表，然后根据该颜色表对原图像的像素点进行聚类分析，产生初始分色图像。再根据C均值聚类的方法对初始聚类中心进行调整，生成新的分色图像，直到满足结束条件后结束算法。实验结果表明，该算法在大大减少原图像的颜色数量的同时基本保持分色图像的质量，是一种实用的分色方法。

An Adaptive Color Segmentation Algorithm Based on Improved C - Mean Clustering

Color image segmentation is widely and importantly applied in textile and printing industry,its aim is to use as less kinds of colors as possible to describe a piece of true color image and make it much similar to original image.This paper presents an adaptive color segmentation algorithm based on improved C-Mean clustering.This algorithm firstly creates a color table at random,and then clusters the pixels in the original image in accordance with the color table,so gets an initial color segmentation image.Secondly,it adjusts the clustering centers by C-Mean clustering and creates new color segmentation image,repeats it until touches the stop condition.The results of experiment indicate that,this method in reducing color kinds of original image,can keep the quality of the color segmentation image at the same time and it is a practical algorithm.