Subcortical structure segmentation (FIRST)
- A Bayesian appearance model that uses deformable models to estimate the boundary of subcortical structures based on extensive, high-quality manual training data. It is also capable of detecting local shape changes via vertex analysis, using non-parametric permutation-based inference (RANDOMISE).
Tissue-type segmentation (FAST)
- Markov Random Field (MRF) implementation of Gaussian mixture model, incorporating simultaneous bias-field estimation and correction.
Brain Extraction (BET)
- A fast and robust method based on a deformable model implementation and local statistics.