一种模糊核聚类的彩色图像量化算法  

A Color Image Quantization Algorithm Based on Fuzzy Kernel Clustering

作　　者：马玉洁[1] 

机构地区：[1]商丘师范学院计算机科学系,河南商丘476000

出　　处：《河南科技大学学报（自然科学版）》2011年第1期45-48,119,共4页Journal of Henan University of Science And Technology:Natural Science

基　　金：河南省自然科学基金项目(092300410220)

摘　　要：提出了一种基于模糊核聚类的彩色图像量化算法。首先用中位切割算法对图像进行初始量化,然后依据NBS距离与人类视觉对颜色差别的定量关系确定初始聚类中心,最后结合模糊核聚类方法对Munsell空间的每个像素进行聚类以实现对颜色的修改,从而完成图像的量化。仿真结果表明:所提算法在量化数目相同的情况下,量化效果明显优于中位切割算法和模糊C均值算法。A color image quantization algorithm based on fuzzy kernel clustering was studied in this paper.Firstly,the original image was quantized by using the median-cut algorithm.Then,the initial clustering centers were determined adaptively on the relationship between the NBS distance and the color difference for human visual.Finally,the clustering center color values were modified by use of the fuzzy kernel clustering algorithm in the Munsell space in order to get the quantization effect.The simulation result shows that the presented algorithm has better quantization effect than the median-cut algorithm and fuzzy c-means algorithm in the same quantization number.

关 键 词：图像 量化 模糊核聚类 中位切割 模糊C均值 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]