Incorporating Nonrigid Registration into Expectation Maximization Algorithm to

Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images By K. M. Pohl, W. M. Wells, A. Guimond, K. Kasai, M. E. Shenton, R. Kikinis, W. E. L. Grimson, and S. K. Warfield Email: pohl@mit. edu MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images

Overview • Introduction • Incorporating Local Prior in EM-MF • Current Implementation – Tools and Tricks • Possible Advancements • Conclusion MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 2

Goal SPGR T 2 W Tissue Atlases MICCAI’ 02 The Magic Automatic Segmenter Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 3

EM-MF Algorithm M-Step Image Correct Intensities Smooth Bias MF-Step Regularize Weights Estimate Tissue Probability E-Step Label Map MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 4

Mean Entropy Atlas MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 5

Merging MEA with SPGR MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 6

Bias MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 7

Bias in Color MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 8

3 D View of SPGR MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 9

Including Local Priors 3. Step Case 1. Step Brain Atlas Registration Probability Maps 2. Step M-Step Correct Bias MF-Step Estimate Align Atlas E-Step Label Map MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 10

Estimating the Tissue Class P(Tissue. TT)| Position [x][y][z]) e. P(Tissue WT[x][y][z] * P(GV[x][y][z] * EM Algorithm | Distribution of T, Bias) e. Energy(WT[x][y][z] | Neighboring W) MF-Approximation Local Prior + GV[x][y][z] = Grey Value at position [x][y][z] WT [x][y][z] = Weights for tissue class T at position [x][y][z] MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 11

Registration only EM-MF Affine EM-MF Non Rigid EM-MF Comparing different Segmenter 2 Channel Input - Segmenting up to 7 tissue classes MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 12

2 Channel Segmentation with Patient Case and 11 Tissue Classes MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 13

Correction of 1 Channel EM-MF-LP through Specialist Background Skin Right/Left Amygdala MICCAI’ 02 CSV White Matter Grey Matter Right/Left Superior Temporalgyrus Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 14

Comparing Manual to EM-MF-LP of the STG Rater B EM-MF-LP Rater A MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 15

Current Installation • Algorithm is a VTK Filter integrated in Slicer • MF Approximation: – Multi Threaded – Lookup Table for Gaussian Distribution • Using several Relaxation Methods instead of the Mean Field Energy Function • Multi Channel Input (SPGR, T 2 , PD) • Train Tissue Definition, e. g. CIM, Distribution • Interface to Matlab MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 16

EM-MF in Slicer MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 17

Tabs of GUI Overview Class Definition Class Interaction EM Settings Skill Level MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 18

Possible Improvements • Registration Step: – After each segmentation re-register case with atlas • E Step – Include shape and topology information in weight calculation – Use local class interaction matrix • M Step: – Use several other filters to smooth bias, e. g. Box Filter, Pascal Triangle, … – Include “trash tissue class” where pixels get assigned if all weights are low Bias does not get corrupted MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 19

Conclusion • Made EM-MF Algorithm more robust • Segmented tissue classes with overlapping gray value distributions • Included spatial/atlas information into E-Step • Cortex pacellation possible • Future Work: Validating Segmentation MICCAI’ 02 Incorporating Non-rigid Registration into Expectation Maximization Algorithm to Segment MR Images 20