融合残差驱动和簇自适应的直觉模糊C均值图像分割算法OA

In view of the problems that the existing intuitionistic fuzzy clustering algorithms have difficulty in determining the optimal number of clustering clusters and are affected by noise points during the clustering process,an IFCM(Intuitionistic Fuzzy C-means)image segmentation algorithm integrating residual-driven and cluster adap-tive is proposed.To alleviate the fuzziness and uncertainty of sample points,based on the intuitionistic fuzzy set framework,both sample membership degree and hesitation degree are considered simultaneously,thereby obtaining a more accurate membership degree.Based on the traditional IFCM,a regularization term of cluster adaptive merging is introduced,enabling the number of clusters to be adaptively adjusted and the optimal number of clusters to be found.This effectively avoids the cumbersome setting of the initial number of clusters and reduces the sensitivity of the ini-tialization parameters.The effectiveness of the proposed algorithm is verified on the artificial image dataset and the magnetic resonance image dataset.The experimental results show that the proposed algorithm can achieve a better noise removal effect while adaptively determining the optimal number of clusters.