Semi-supervised clustering for MR brain image segmentation

作者：

Highlights：

• We propose a semi-supervised clustering method for MR brain image segmentation.

• The initial parameters are calculated based on prior knowledge.

• The proposed method accelerates the convergence of the segmentation.

• Performance and accuracy gain observed in the experiments.

摘要

•We propose a semi-supervised clustering method for MR brain image segmentation.•The initial parameters are calculated based on prior knowledge.•The proposed method accelerates the convergence of the segmentation.•Performance and accuracy gain observed in the experiments.

论文关键词：Semi-supervised learning,Gaussian mixture model,Contextual segmentation,Magnetic resonance brain multispectral image

论文评审过程：Available online 4 September 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.08.046