Clustering of magnetic resonance images
Geoffrey John McLachlan, Shu‐Kay Ng, Graham J. Galloway, Diping Wang · 1996
We consider a statistical model-based approach to the clustering of multispectral magnetic resonance (MR) images. The method of maximum likelihood is implemented via the EM algorithm, using an approximation to the E-step based on the ICM algorithm. The approach is illustrated on some MR images of the human brain.