Tissue Mixture Characterization In The Presence of Mri Inhomogeneity by the Em Algorithm
Z. Liang, L. Li, Daria Eremina, Hongbing Lu · 2006
This paper presents a model-based approach to correct for both partial volume effect and inhomogeneity in segmenting tissue mixtures inside each voxel of magnetic resonance images. A maximum a posteriori probability (MAP) solution is sought. In calculating the solution, the well-known expectation maximization (EM) algorithm is employed. The models of data likelihood and Markov priors for tissue mixture and bias field in establishing this MAP-EM framework are described in details. A preliminary test is presented