A Robust Fuzzy c-Means Clustering Model with Spatial Constraint for Brain Magnetic Resonance Image Segmentation

摘要

Due to the effect of various objective factors, there is often noise in the brain magnetic resonance image. Although fuzzy c-means clustering is an efficacious segmentation method, it is quite sensitive to some noises or outliers in image. In order to improve the segmentation accuracy and robustness to noise, a robust fuzzy c-means clustering model with spatial constraint is proposed for unsupervised segmentation of brain magnetic resonance images in this paper. The spatial distance and the gray level information between the local neighborhood pixels are combined with the non-linear weighting form in the similarity measure of the fuzzy clustering, the constrained relationship between the center pixel and its neighboring pixels can be more accurately described in the local region, and the objective function is more reasonably established. In addition, it is helpful to improve the clustering performance by introducing the local neighborhood information into the fuzzy membership degree. The algorithm is implemented in brain magnetic resonance images that the experimental results demonstrated that the proposed algorithm is more robust to noise and effectively preserves the image detail than fuzzy c-means clustering algorithm and its variants.