Machine learning: Expectation-Maximization, Random Forests
Statistical modeling: Bayesian inference, Maximum Likelihood Estimators, MAP problems
Fig1. - Axial, sagital and coronal views of segmented flair hyperintensities acquired from a patient suffering from vanishing white matter disease. Segmentation was calculated based on histogram fitting: a gaussian was used to fit grey and white matter voxels and a frechet distribution (an extreme value distribution) was used to fit the hyperintensities.
Brain tumor growth modeling
Temporal graphical models
Markov random fields
Energy minimization
Uncertainty estimation
Patient-specific characterization of Glioblastoma Multiforma
Precision medicine: Personalized glioma modeling has become necessary due to intertumoral cellular and molecular heterogeneity
Radiogenomics: Investigating the relation between genetic profiles and MRI phenotypes