Kernel-based fuzzy local information clustering algorithm self-integrating non-local information

The application of fuzzy clustering in image segmentation is a research hotspot nowadays. Existing robust have some problems such as the inability to adaptively select spatial constraint parameters, accurately segment images corrupted by high noise, and achieve balance between noise suppression detail preservation. In based on objective function optimization, choice distance measure very important. Gaussian kernel defined Euclidean has been widely used many fields pattern recognition, machine learning, etc. However, sensitive outliers or it difficult obtain satisfactory results for special data sets, which will affect performance algorithm. this paper, non local information self-integration optimization algorithm kernel-based proposed. uses method basis introduces non-local at same time, solves common current Firstly, self-integrating problem selecting continues self-learning iteratively calculates parameter values; secondly, induced further enhance robustness against adaptability processing sets; Finally, are integrated time effect, can effectively suppress most retain details original image. Experimental show that proposed superior existing state-of-the-art clustering-related presence noise.