基于颜色空间和优化初始中心的模糊c均值聚类算法

FUZZY c MEANS CLUSTERING ALGORITHM FOR COLOR SPACE AND OPTIMIZED INITIAL CENTER

提出了一种基于HIS空间的优化初始中心的模糊c-均值的彩色图像分割方法．首先将彩色图片由RGB转换为HIS，并将H和1分开处理，通过计算样本的权重，选取有代表性的样本作为初始聚类中心，给出优化初始聚类中心的FCM算法，将该算法应用于H和I通道，得出新的基于颜色空间的FCM算法．该算法可以得到较稳定的结果，并且提高了聚类的准确率．

A color image segmentation method to optimize the initial center in HIS color space based on the fuzzy C - mean. Firstly, the color image is transformed to HIS from RGB, and the H and I component is separated. By computing samples＇ weight, the representative samples are selected as the initial cluster center. Then the FCM algorithm based on the optimize initial center is gotten. The application of this algorithm in H and I component obtains a new algorithm FCM based on color space. This algorithm can obtain more stable resuhs, and improves the accuracy of clustering.