基于偏移场校正的模糊局部C-均值聚类算法

Fuzzy C-means Clustering Algorithm Based on the Bias Field Correction

为了提高灰度不均匀医学图像分割性能,提出了一种基于偏移场校正的模糊局部C-均值聚类算法。该算法将含偏移场校正的医学图像距离模型应用于模糊局部C-均值聚类算法,并对得到聚类目标函数进行严格推导,获得迭代求解的隶属度、聚类中心和偏移场表达式,来构造聚类分割算法。对比实验表明,相对于现有分割算法,该分割算法在处理医学图像时能够获得较好的分割效果。

In order to improve the segmentation performance of gray uneven medical image,a fuzzy C-means clustering algorithm based on the bias field correction is proposed in this paper. This algorithm applies the medical image distance model which contains bias field correction to fuzzy local information C-means clustering algorithm. The iterative expression of degree of membership,clustering center and bias field are obtained through strict mathematical derivation of the clustering segmentation objective function. Compared with the existing fuzzy C-means clustering segmentation algorithm,this algorithm is more effective on medical image segmentation.