烟花算法优化的软子空间MR图像聚类算法

发布时间：2018-04-19 16:17

本文选题：烟花算法 + 软子空间聚类　；参考：《软件学报》2017年11期

【摘要】：现有的软子空间聚类算法在分割MR图像时易受随机噪声的影响,而且算法因依赖于初始聚类中心的选择而容易陷入局部最优,导致分割效果不理想.针对这一问题,提出一种基于烟花算法的软子空间MR图像聚类算法.算法首先设计一个结合界约束与噪声聚类的目标函数,弥补现有算法对噪声数据敏感的缺陷,并提出一种隶属度计算方法,快速、准确地寻找簇类所在子空间;然后,在聚类过程中引入自适应烟花算法,有效地平衡局部与全局搜索,弥补现有算法容易陷入局部最优的不足.EWKM,FWKM,FSC,LAC算法在UCI数据集、人工合成图像、Berkeley图像数据集以及临床乳腺MR图像、脑部MR图像上的聚类结果表明,所提出的算法不仅在UCI数据集上能够取得较好的结果,而且对图像聚类也具有较好的抗噪性能,尤其是对MR图像的聚类具有较高的精度和鲁棒性,能够较为有效地实现MR图像的分割.
[Abstract]:The existing soft subspace clustering algorithms are susceptible to random noise in the segmentation of Mr images, and the algorithm is prone to fall into local optimum because of its dependence on the selection of initial clustering centers, which leads to unsatisfactory segmentation results.To solve this problem, a soft subspace Mr image clustering algorithm based on fireworks algorithm is proposed.The algorithm first designs an objective function combining bound constraints and noise clustering to make up for the shortcomings of the existing algorithms which are sensitive to noise data, and proposes a method for calculating membership degree, which can quickly and accurately find the subspace where the cluster is located.In the process of clustering, adaptive fireworks algorithm is introduced to balance the local and global search effectively, to make up for the deficiency of the existing algorithms. EWKM / FWKMU / FSCSC-LAC algorithm is applied to UCI data sets, artificial synthetic images, Berkeley image datasets and clinical mammary Mr images.The clustering results on brain Mr images show that the proposed algorithm can not only obtain better results on UCI datasets, but also have a good anti-noise performance for image clustering.Especially, it has high accuracy and robustness to the clustering of Mr images, and it can effectively realize the segmentation of Mr images.
【作者单位】：陕西师范大学计算机科学学院;中国科学院深圳先进技术研究院生物医学与健康工程研究所;
【分类号】：TP18;TP391.41

【相似文献】