STAT%20992:%20Streamlined%20Image%20Analysis%20Framework

1
STAT 992 Streamlined Image Analysis
Framework Moo K. Chung mchung_at_stat.wisc.edu
Topics mentioned briefly
Topics covered
2. Reading Image into computer Image formats
ascii, binary DIACOM, MINC, ANALYZER, BRIK
1. Image collection MRI, fMRI, PET, CT, EEG, MEG
Image Acquisition
MLE LSE Inverse problem
Topics covered In detail
3. Gaussian mixture EM algorithm and a Bayesian
prior
4. Deformable curve and surface model Active
contour, snake
5. PDF curve evolution Level set, 0-levelset of
spline
Image Segmentation
E-step, Jensens inequality
6. Similarity measure Intensity correlation,
mutual information Kullback-Leibler divergence
SPM99, AIR package, Tensor-based
morphometry, Displacement modeling
Image Registration
7. Curve registration dynamic time warping,
iterative shifting
8. Image registration Intensity-based, PDE-based
Functional data Analysis, Time warping
9. Random fields Mean-square convergence Hilbert
space Karhunen-Loeve expansion White noise,
Brownian motion
10. Linear additive model General linear model
(GLM) Spatially correlated noise modeling on
manifolds
Statistical Modeling
Covariance function estimation
Mixed effect model fMRI time series
Geometry Modeling
11. Curve modeling Curvature estimation, Riemannia
n metric tensor
12. Surface modeling Curvature estimation, Riemann
ian metric tensor
Normal vector estimation, Regression on
manifolds Asymmetry analysis
13. Kernel Smoothing Fourier transform
14. Diffusion Smoothing Diffusion
equations, Spherical harmonics, Finite element
method
16. Directional smoothing Edge preserving
smoothing Vector fields smoothing
Image Smoothing
Voxel-based morphometry Nadaraya-Watson estimator
QQplot Histogram Reduction of false-positives
15. Heat kernel smoothing Iterative kernel
smoothing, Laplace-Beltrami operator
Weighted polynomial regression, Vector-spline,
Tensor field smoothing Diffusion tensor imaging
Parametrix expansion, Geodesics M-estimation,
M-smoothing
18. Random fields theory Corrected P-value Euler
characteristic
Multiple comparisons
19. Permutation test Checking image processing
artifacts
17. False discovery rate (FDR) Combining with RFT
20. Statistical Inference Local, global inference
Inference on manifolds
Correlation analysis requires correlation random
fields.
Optimal bandwidth as P-value minimizer