基于空间模糊核聚类的脑肿瘤图像分割方法

发布时间：2018-12-06 13:52

【摘要】：针对传统FCM算法对脑肿瘤图像分割时对噪声敏感且没有考虑图像空间信息出现分割结果不准确的问题,提出一种新的基于空间模糊核聚类(Spatial Kernel Fuzzy C-means,SKFCM)的脑肿瘤MR图像分割方法。首先,算法引入核函数,将数据由低维特征空间转化为高维特征空间,提高了不同类别特征差异。其次,根据图像目标区域梯度差异,对KFCM算法进行改进,将图像空间信息引入到KFCM算法,用目标区域方差总和与边界梯度均值倒数之和作为新的目标函数。最后,对脑肿瘤图像的样本特征进行迭代优化,实现了脑肿瘤的精确分割。算法分割精确度得到了提高,并且通过图像的空间信息增加了方法的鲁棒性。实验结果表明,改进后算法对脑肿瘤MR图像不同类别间区分度高且具有较强噪声抑制能力,实现了较好的分割效果。
[Abstract]:Aiming at the problem that the traditional FCM algorithm is sensitive to noise and does not consider the inaccurate segmentation results of image spatial information, a new method based on spatial fuzzy kernel clustering (Spatial Kernel Fuzzy C-means) is proposed. MR image segmentation method of brain tumor based on SKFCM. Firstly, the kernel function is introduced to transform the data from low-dimensional feature space to high-dimensional feature space, which improves the difference of different kinds of features. Secondly, according to the gradient difference of the target region, the KFCM algorithm is improved, and the spatial information of the image is introduced into the KFCM algorithm. The sum of the total variance of the target region and the reciprocal of the mean of the boundary gradient is used as the new objective function. Finally, the sample feature of brain tumor image is optimized iteratively, and the accurate segmentation of brain tumor is realized. The segmentation accuracy of the algorithm is improved, and the robustness of the method is enhanced by the spatial information of the image. The experimental results show that the improved algorithm has a high degree of discrimination and strong noise suppression ability among different categories of brain tumor MR images, and achieves a better segmentation effect.
【作者单位】：内蒙古科技大学信息工程学院;
【基金】：国家自然科学基金项目(61179019)资助内蒙古自然基金面上项目(2015MS0604)资助内蒙古自治区高等学校科学研究项目(NJZY145)资助
【分类号】：R739.41;TP391.41

【相似文献】