一种结合邻域信息的 FCM 图像分割新算法

摘要

提出一种结合邻域信息的 FCM 图像分割的新算法。将空间邻域信息引入隶属度函数，改进空间函数，并对隶属度进行上下截集半模糊化处理，最后修正隶属度的硬划分矩阵，提高了图像分割质量。实验结果表明，新算法在增强算法抗噪性的同时能够更有效地进行图像分割。%In this paper,we present a new Fuzzy C - means(nSFCMp )algorithm incorporating the spatial in-formation for image segmentation,which proposes a new spatial function,and incorporates spatial informa-tion into the membership function. In addition,the value of up and down cut - off is applied to half fuzz membership matrix that enhances the convergence speed of algorithm. What’s more,modifying the member-ship’s hard partition matrix is effective to improve the quality of segmentation. Experimental result shows that the new algorithm is more robust to noise and efficient in image segmentation.