Title: Overview of Medical
1Overview of Medical
Image Registration
J. Michael Fitzpatrick, Department of Electrical
Engineering and Computer ScienceVanderbilt
University, Nashville, TN CS/EECE 359, Fall,
2007
2Acknowledgements
- Benoit M. Dawant, PhD, EECS
- Robert L. Galloway, PhD, BME
- William C. Chapman, MD, Surgery
- Jeannette L. Herring, PhD, EECS
- Jim Stefansic, PhD, Psychology
- Diane M. Muratore, MS, BME
- David M. Cash, MS, BME
- Steve Hartman, MS, BME
- W. Andrew Bass, BME
NSF NIH
Matthew Wang, PhD, IBM Jay B. West, PhD, Accuray,
Inc. Derek L. G. Hill, PhD Kings CollegeCalvin
R. Maurer, Jr., PhD, Stanford U.
3Computed Tomography (1972)
Siemens CT Scanner (Somatom AR)
43D Cross-sectional Image
voxels (volume elements)
5Magnetic Resonance Imaging
GE MR Scanner (Signa 1.5T)
6Positron Emission Tomography
GE PET Scanner
7Physician has 3 or more views.
MR (wet tissue)
PET (biologicalactivity)
CT (bone)
8Combining multiple images requires image
registration
9Image Registration Definition
Determination of corresponding points in two
different views of the same object
10Motion relative to the scanners can be
three-dimensional.
11Slice orientations vary widely.
transverse
sagittal
coronal
12Views may be very different.
13But all orientations and all views can be
combined if we have the 3D point mapping.
14Combining Registered Images Image Fusion
15Rigid Registration Definition
- Rigid Registration Registration using a rigid
transformation
16Rigid Transformation
Distances between all points remain constant.
Rigid
Non-rigid
17Nonrigid Transformationscan be very complex!
Thompson, 1996
18Registration Methods
- Retrospective methods
- Match anatomical features e.g., surfaces
- Maximize similarity of intensity patterns
- Prospective methods
- Non-invasive Match skin markers
- Invasive Match bone-implanted markers
19Mutual Information An Example of
Intensity-Based Registration
Also known as Voxel-Based Registration
202D Intensity Histogram (Hill94)
21Misregistration Blurs It
5 cm
0 cm
2 cm
MR
CT
MR
PET
Hill, 1994
22Mutual Information(Viola, Collignon, 1996)
- A measure of histogram sharpness
- Most popular intensity method
- Assumes a search method is available
- Stochastic, multiresolution search common
- Requires a good starting pose
- May not find global optimum
23Example Mutual Information
Studholme, Hill, Hawkes, 1996, Automated 3D
registration of MR and CT images of the head,
MIA, 1996
24Goal for 1st Day
25The Iterative Closest-Point AlgorithmAn
Example of Surface-Based Registration
26Iterative Closest-Point Method(Besl and McKay,
1992)
- Minimizes a positive distance function
- Most popular surface method
- Assumes surfaces have been delineated
- Guaranteed to converge
- Requires a good starting pose
- May not find global optimum
27Start with two surfaces
28Reorient one (somehow)
29Reorient one (somehow)
30Reorient one (somehow)
31Pick points on moving surface
32Pick points on moving surface
33Remove moving surface
34Points become proxy for surface
35Find closest points on stationary surface
36Got here 10/4/2005
37Measure the total distance
38Remove stationary surface
39Points become proxy for surface
40Register point sets (rigid)
41Register point sets (rigid)
42Restore stationary surface
43Find (new) closest points
44Find (new) closest points
45Remove stationary surface
46Remove stationary surface
47Register Points
48Register Points, and so on
49Iterative Closest-Point Algorithm
- Find closest points
- Measure total distance
- Register points
50Example ICP for Head
Dawant et al.
51Example ICP for Vertebra
Muratore, Herring, Dawant, Galloway
52Got to here in CS 395, 8/28/03
53ICP requires surface delineation, which is a
problem in Image Segmentation
Example Level Set Segmen-tation (Dawant et al.)
http//www.vuse.vanderbilt.edu/dawant/levelset_ex
amples/
54(1) Point-Based Registration (2) Prospective
Methods(3) Image-guided Therapy
The Fiducial Marker An Example of
55Image-Guided Surgery
Just another image registration problem.
- ...and the other is the patient.
56AcustarAllen, Maciunas, Fitzpatrick, and
Galloway1986-1996 (JJ ? Z-Kat)
Posts
are implanted into the skull.
Maurer, et al., TMI, 1997
57AcustarAllen, Maciunas, Fitzpatrick, and
Galloway1986-1996 (JJ ? Z-Kat)
Liquid in marker shows up in image
Maurer, et al., TMI, 1997
58AcustarAllen, Maciunas, Fitzpatrick, and
Galloway1986-1996 (JJ ? Z-Kat)
Marker center and cap center occupy the same
position relative to the post
Maurer, et al., TMI, 1997
59AcustarAllen, Maciunas, Fitzpatrick, and
Galloway1986-1996 (JJ ? Z-Kat)
Marker center and cap center occupy the same
position relative to the post
Maurer, et al., TMI, 1997
60Point-based, Rigid Registration
61What to Optimize?
- Mean-square Fiducial Registration Error (FRE2)
- Known as the Orthogonal Procrustes Problem in
statistics since 1950s. - Robust estimators (median, M-estimators)
- Less sensitive to outliers
Color key Major problems solved, Negligible
work
62Minimization of FRE2(Shönemann, Farrell, 1966)
- Minimizes a positive distance function
- Most popular point method
- Assumes points have been localized
- Guaranteed to converge
- Does not require a good starting pose
- Always finds global optimum
63Sum of Squares Step 1
Center the points
64Step 2 (Shönemann, Farrell, 1966)
Centered
Determine the Rotation
Centered and Rotated
65Finding Points Localization
66Competition Acustar v. Leibinger
67Competition Acustar v. Leibinger
68Accuracy
69Measures of Registration Error
View 2
View 1
Registered Views
70Acustar, 3mm Slice CT-physical TRE 0.3 to
0.7 mm
71Registration of Head Images The State of the Art
Retrospective Median Maximum
(Acustar) Best CT-MR 0.6 mm 3.0 mm
(0.5 mm) Poor CT-MR 5.4 mm 61 mm
(0.5 mm) Best PET-MR 2.5 mm 6.0
mm (1.7 mm) Poor PET-MR 5.3 mm
15 mm (1.7 mm)
And how do we know?
72Retrospective Image Regstration Evaluation
1995-2007
External site
Access 150 participants in 20
countries Evaluation 57 participants in 17
countries
Vanderbilt
73Goal for 2nd Day
74Error Theory for Minimization ofMean-square FRE
75Start with Assumptions about FLE
Independent, normal, isotropic, zero mean
76Effective FLE
Space 2
Space 1
77FRE Statistics Sibson 79
Approximate Solution
78FRE Statistics Sibson 1979
Approximate Solution
79TRE statistics, 1998
Approximate Solution
Fitzpatrick, West, Maurer, TMI, 98
Configuration doesmatter.
Principal axes
80TRE forFLEof1mm
Marker Placement
West et al., Neurosurgery, April, 2001
FRE 1mm
81A distribution would be better
ltTRE2gt
Probability density
TRE2
82And what about direction?
83TRE statistics, 2001
Approximate Solution
TRE1
TRE3
TRE2
West and Fitzpatrick., TMI, Sep 2001
84Some Remaining Problems
85Isotropic Scaling
Actually now solved Batchelor, West,
Fitzpatrick, Proc. of Med. Im. Undstnd. Anal.
,Jul 2002
86Anisotropic Scaling
(Iterative Solution Only)
87Register M points sets simultaneously
View 1
View 2
The Generalized Procrustes Problem
View 3
View M
(Iterative Solution Only)
88Spatial Weighting
(Iterative Solution Only)
89Other Unsolved Problems
- What is the statistical effect on TRE of dropping
or adding a fiducial? - Does anisotropy in FLE always, sometimes, or
never makes TRE worse? - How do we configure markers on a given surface so
as to minimize TRE over a given region? - Is there a correlation between FRE and TRE?
90Other Unsolved Problems (cont.)
- Extenion to perspective transformations.
- Extension to surface matching.
91End of TalkAdditional slides follow
92Categories within error prediction
- Number of point sets Two or more
- Scaling Isotropic or anisotropic
- Point-wise weighting equal or unequal
- Anisotropic weighting
- Cost function squared error or other
- Point-wise FLE equal or unequal
- Spatial FLE isotropic or anisotropic...
Key Approximate, Negligible progress
93Anisotropic Scaling
Iterative Algorithm
Problem Statement
94Scaling Anisotropic II
Iterative Algorithm
Problem Statement
R, t rotation, translationwi2 point
weighting S diag( sx , sy , sz )
Given xi yi wi find R, t, S to minimize
mean FRE2
95Spatial Weighting
Iterative Algorithm
Problem Statement
Partial Solution
96Generalized Procrustes Problem
Cost function
Iterative method(only)
97Add Isotropic Scaling
Approximate Solution
FRE2 sum of squared fiducial registration errors
98FRE Generalized Scaling
Approximate Solution
FRE2 sum of squared fiducial registration errors
99TRE statistics with scaling
Approximate Solution
TRE2 target registration error
100Applications of TREStatistics
101Error Bounds
102Probe Design
103Fiducial-Specific FRE
Poor fiducial alignment tends to occur where
target registration is good!!
104Four Solution Methods
( All work equally well Eggert91! )
105Generalized Procrustes Problem
(Weve already done it for M2.)
Problem Statement
Illustration
106Generalized Procrustes Problem
Iterative Algorithm
Illustration
Subject to S(m) normalization
107Approximation Method
(due to Sibson, 1979)
108Approximation Method (cont.)
109FRE Statistics
Approximate Solution
Problem Statement
110TRE statistics with scaling
Approximate Solution
Problem Statement
111What do solved and unsolved mean?
- Solved, working definition
- Reduced to solving algebraic equations
- Iterative algorithm that converges to solution
- Approximate solution accurate to
- Unsolved
- Not solved
112Point-wise weighting Equal or Unequal
(Weve just looked at this one.)
Problem Statement
Solution
See previousslides again!
1131. Performing a Registration
a.k.a. The Orthogonal Procrustes Problem
Problem Statement
1142. Predicting Registration Error
- Input---
- fiducial positions
- target position, r
- FLE distribution
View 2
View 1
Registered Views
115Isotropic Scaling
Problem Statement
Solution