图像分割的EMKPFC算法

The EMKPFC algorithm of image segmentation

现有基于模糊聚类的图像分割算法对噪声敏感,不能妥善地处理图像的灰度特征与邻域像素之间关系.针对该问题,在可能性聚类的基础上融入多核聚类思想,提出了图像分割的EMKPFC算法(Enhanced Multiple Kernel Possibilistic Fuzzy C-means algorithms).该算法可以有效地利用模糊聚类方法以及可能性聚类算法的优点.进一步地,该算法能够规避普通核算法对于核函数选择的不确定性,增加了算法的抗变换性;对于挑选的多种核函数,凭借权重组合能够满足不同图像对于各种核函数的偏好需求,计算出最佳匹配的权重值.在没有任何先验的情况下,不仅可以进行准确的划分,而且还可以做到划分非线性团状样本.通过对于人造图像、真实图像和医学图像的实验结果表明,所提算法比其他相关基于模糊聚类的图像分割算法都具有更好的效果.

Current image segmentation algorithm based on fuzzy clustering is sensitive to noise and cannot properly deal with the relationship between the gray features of the image and the neighborhood pixels. To solve this problem,the idea of multiple kernel clustering and the possibility of clustering are introduced to propose the EMKPFC（Enhanced Multiple Kernel Possibilistic Fuzzy C-means）algorithm for image segmentation. This algorithm can effectively combine the advantages of the fuzzy clustering method and the possibility clustering algorithm. Fur- thermore, the algorithm can avoid the uncertainty of the selection of kernel function and increase the ability of anti- transformation. For a variety of chosen kernel functions, it can meet the different demand of images for nuclear functions and calculate the best match of the weight value. In the case of the data without any priori,it is not only able to accurately divide the leading data, but also can be done with the non-linear classification of non group data. The results of experiments on synthetic images, real images and medical images show that the proposed algorithm has better performance than other image segmentation algorithms.