Robust spatially constrained fuzzy c-means algorithm for brain MR image segmentation

作者：

Highlights：

• A robust spatially constrained fuzzy c-means algorithm is proposed.

• The negative log-posterior is utilized as dissimilarity function.

• A novel factor is proposed by considering the spatial direction.

• The factor is calculated based on the posterior and prior probabilities.

• Our algorithm can substantially improve the accuracy of brain MR image segmentation.

摘要

•A robust spatially constrained fuzzy c-means algorithm is proposed.•The negative log-posterior is utilized as dissimilarity function.•A novel factor is proposed by considering the spatial direction.•The factor is calculated based on the posterior and prior probabilities.•Our algorithm can substantially improve the accuracy of brain MR image segmentation.

论文关键词：Image segmentation,Magnetic resonance imaging,Fuzzy c-means,Spatial information,Intensity inhomogeneity

论文评审过程：Received 21 June 2013, Revised 12 October 2013, Accepted 27 January 2014, Available online 4 February 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.01.017