【摘要】 基于二维直方图的模糊聚类分割方法,可以有效地抑制噪声。但是FCM(模糊C均值)算法用于图像聚类时最大的缺点是运算开销太大,进而限制了该算法在图像分割中的应用。通过构造合理的二维直方图,并筛选出符合规定条件的元素作为聚类样本,再结合FCM算法进行图像分割。实验结果表明该方法具有与基于一维直方图的模糊聚类分割方法速度相近,但却比其分割精度高很多的良好特点。更多还原

【Abstract】 Fuzzy clustering algorithm based on 2D histogram can suppress noise effectually.But Fuzzy C-Means(FCM) clustering algorithm has a large disadvantage,that is long run time,which restricts its application in the image segmentation.By constructing the reasonable 2D histogram and selecting the elements tallying with given condition,this paper studies image segmentation based on FCM algorithm.The experimental results illustrate that the proposed algorithm costs approximate same time with the algorithm based on 1D histogram,but prior to the latter such as high segmentation precision.更多还原