摘要
在对手抑制式模糊C-均值聚类算法中,参数α的选择有可能导致原有的隶属度之间顺序的改变。针对其不足,提出了一种改进的模糊C-均值聚类算法,他是通过引入2个不同的调节参数1α和2α修正不同大小的隶属度,在保持隶属度的次序不变的前提下可以加速图像分割的收敛速度。实验表明,该算法不但能有效地提高聚类的速度,且能得到较好的分割效果。
To rival checked fuzzy C- means clustering algorithm, the choice of parameter a is possible to change the sequence of the primary degree of membership. To overcome defect of the algorithm,a new improved algorithm is proposed. The new improved algorithm corrects the degree of membership by two differernt parameters α1 and α2 . It can accelerate the convergent speed while keeeping the sequence of the membership. Experiment results show the algorithm can not only improve the speed, but also can get the better effect of segmentation.
出处
《现代电子技术》
2007年第13期180-181,共2页
Modern Electronics Technique
