Computed TomographyComputed Tomography (part...
Transcript of Computed TomographyComputed Tomography (part...
![Page 1: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/1.jpg)
Computed TomographyComputed Tomography(part I)
Yao WangPolytechnic University Brooklyn NY 11201Polytechnic University, Brooklyn, NY 11201
Based on J. L. Prince and J. M. Links, Medical Imaging Signals and Systems, and lecture notes by Prince. Figures are from the textbook.y , y g
![Page 2: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/2.jpg)
Lecture Outline• Instrumentation
– CT Generations– X-ray source and collimation– CT detectors
• Image Formation• Image Formation– Line integrals– Parallel Ray Reconstruction
• Radon transform• Back projection• Filtered backprojection• Convolution backprojection• Implementation issues
EL5823 CT-1 Yao Wang, NYU-Poly 2
![Page 3: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/3.jpg)
Limitation of Projection Radiography• Projection radiography
– Projection of a 2D slice along one direction only– Can only see the “shadow” of the 3D body
• CT: generating many 1D projections in different anglesWhen the angle spacing is sufficiently small can reconstruct the– When the angle spacing is sufficiently small, can reconstruct the 2D slice very well
EL5823 CT-1 Yao Wang, NYU-Poly 3
![Page 4: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/4.jpg)
1st Generation CT: Parallel Projections
EL5823 CT-1 Yao Wang, NYU-Poly 4
![Page 5: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/5.jpg)
2nd Generation
EL5823 CT-1 Yao Wang, NYU-Poly 5
![Page 6: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/6.jpg)
3G: Fan Beam
Much faster than 2G
EL5823 CT-1 Yao Wang, NYU-Poly 6
![Page 7: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/7.jpg)
4G
FastCannot use collimator at detector, hence affected by scatteringscattering
EL5823 CT-1 Yao Wang, NYU-Poly 7
![Page 8: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/8.jpg)
5G: Electron Beam CT (EBCT)
EL5823 CT-1 Yao Wang, NYU-Poly 8
![Page 9: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/9.jpg)
6G: Helical CT
EL5823 CT-1 Yao Wang, NYU-Poly 9
![Page 10: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/10.jpg)
7G: Multislice
EL5823 CT-1 Yao Wang, NYU-Poly 10
![Page 11: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/11.jpg)
X-ray Source
EL5823 CT-1 Yao Wang, NYU-Poly 11
![Page 12: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/12.jpg)
X-ray Detectors
EL5823 CT-1 Yao Wang, NYU-Poly 12
![Page 13: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/13.jpg)
CT Measurement Model
EL5823 CT-1 Yao Wang, NYU-Poly 13
![Page 14: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/14.jpg)
EL5823 CT-1 Yao Wang, NYU-Poly 14
![Page 15: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/15.jpg)
CT Number
Need 12 bits to represent
EL5823 CT-1 Yao Wang, NYU-Poly 15
Need 12 bits to represent
![Page 16: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/16.jpg)
Parameterization of a LineEach projection line is defined by (l,)
sy A point on this line (x,y) can
be specified with two options
x
Option 1 (parameterized by s):
= l
Option 2:
EL5823 CT-1 Yao Wang, NYU-Poly 16
![Page 17: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/17.jpg)
Line Integral: parametric form
EL5823 CT-1 Yao Wang, NYU-Poly 17
![Page 18: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/18.jpg)
Line Integral: set form
EL5823 CT-1 Yao Wang, NYU-Poly 18
![Page 19: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/19.jpg)
Physical meaning of “f” & “g”
EL5823 CT-1 Yao Wang, NYU-Poly 19
![Page 20: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/20.jpg)
What is g(l,)?
EL5823 CT-1 Yao Wang, NYU-Poly 20
![Page 21: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/21.jpg)
Example• Example 1: Consider an image slice which contains a
single square in the center. What is its projections along 0, 45, 90, 135 degrees?
• Example 2: Instead of a square, we have a rectangle. RepeatRepeat.
EL5823 CT-1 Yao Wang, NYU-Poly 21
![Page 22: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/22.jpg)
Sinogram
EL5823 CT-1 Yao Wang, NYU-Poly 22
![Page 23: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/23.jpg)
Backprojection• The simplest method for reconstructing an image from a
projection along an angle is by backprojectionAssigning every point in the image along the line defined by (l ) the– Assigning every point in the image along the line defined by (l,) the projected value g(l, ), repeat for all l for the given
sy
x
EL5823 CT-1 Yao Wang, NYU-Poly 23
![Page 24: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/24.jpg)
EL5823 CT-1 Yao Wang, NYU-Poly 24
![Page 25: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/25.jpg)
Two Ways of Performing Backprojection• Option 1: assigning value of g(l, ) to all points on the line (l, )
– g(l, ) is only measured at certain l: ln=n l – If l is coarsely sampled (l is large) many points in an image will not be– If l is coarsely sampled (l is large), many points in an image will not be
assigned a value– Many points on the line may not be a sample point in a digital image
• Option 2: going through all sampling points (x,y) in an image, find its p g g g p g p ( y) gcorresponding “l=x cos y sin ” for the given , take the g value for (l, )
(l ) i l d t t i l l l– g(l, ) is only measured at certain l: ln=n l– must interpolate g(l, ) for any l from given g(ln, )
• Option 2 is better, as it makes sure all sample points in an image are assigned a valueassigned a value
• For more accurate results, the backprojected value at each point should be divided by the length of the underlying image in the projection direction (if known)
EL5823 CT-1 Yao Wang, NYU-Poly 25
p j ( )
![Page 26: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/26.jpg)
Backprojection Summation
Replaced by a p ysum in practice
EL5823 CT-1 Yao Wang, NYU-Poly 26
![Page 27: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/27.jpg)
Implementation Issues
EL5823 CT-1 Yao Wang, NYU-Poly 27
From L. Parra at CUNY, http://bme.ccny.cuny.edu/faculty/parra/teaching/med-imaging/lecture4.pdf
![Page 28: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/28.jpg)
Implementation: Projection• To create projection data using computers, also has similar problems.
Possible l and q are both quantized. If you first specify (l,q), then find (x,y) that are on this line. It is not easy. Instead, for given q, you can go through all (x y) and determine corresponding l quantize l to one of those you wantall (x,y) and determine corresponding l, quantize l to one of those you want to collect data.
• Sample matlab code (for illustration purpose only)[x,y]=meshgrid([0:J-1]-J/2,[0:I-1]-I/2);N=ceil(sqrt(I*I+J*J))+1;N ceil(sqrt(I I J J)) 1;N0= floor((N-1)/2);ql=1;G=zeros(N,180);for phi=0:179f ( /2 /2 1 /2 /2 1)for (x=-J/2:J/2-1; y=-I/2:I/2-1)
l=x*cos(phi*pi/180)+y*sin(phipi/180);l=round(l/ql)+N0+1;If (l>=1) && (l<=N)
G(l phi+1)=G(l phi+1)+f(x+J/2+1 y+I/2+1);G(l,phi+1) G(l,phi+1)+f(x+J/2+1,y+I/2+1);Endend
end
EL5823 CT-1 Yao Wang, NYU-Poly 28
![Page 29: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/29.jpg)
Example• Continue with the example of the image with a square in
the center. Determine the backprojected image from each projection and the reconstruction by summing different number of backprojections
EL5823 CT-1 Yao Wang, NYU-Poly 29
![Page 30: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/30.jpg)
Problems with Backprojection Blurring
EL5823 CT-1 Yao Wang, NYU-Poly 30
![Page 31: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/31.jpg)
Projection Slice Theorem
dlljlgG }2exp{),(),(
The Fourier Transform of a projection at angle is a line in the Fourier transform of the image at the same angle.If (l ) are sampled sufficiently dense then from g (l ) we
EL5823 CT-1 Yao Wang, NYU-Poly 31
If (l,) are sampled sufficiently dense, then from g (l,) we essentially know F(u,v) (on the polar coordinate), and by inverse transform can obtain f(x,y)!
![Page 32: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/32.jpg)
Illustration of the Projection Slice TheoremTheorem
EL5823 CT-1 Yao Wang, NYU-Poly 32
![Page 33: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/33.jpg)
Proof• Go through on the board• Using the set form of the line integral• See Prince&Links, P. 198
dlljlgG }2exp{)()(
dlljlgG }2exp{),(),(
EL5823 CT-1 Yao Wang, NYU-Poly 33
![Page 34: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/34.jpg)
The Fourier Method• The projection slice theorem leads to the following
conceptually simple reconstruction method– Take 1D FT of each projection to obtain G() for all – Convert G() to Cartesian grid F(u,v)– Take inverse 2D FT to obtain f(x,y)Take inverse 2D FT to obtain f(x,y)
• Not used because– Difficult to interpolate polar data onto a Cartesian grid– Inverse 2D FT is time consuming
• But is important for conceptual understanding– Take inverse 2D FT on G( ) on the polar coordinate leads toTake inverse 2D FT on G() on the polar coordinate leads to
the widely used Filtered Backprojection algorithm
EL5823 CT-1 Yao Wang, NYU-Poly 34
![Page 35: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/35.jpg)
Filtered Backprojection• Inverse 2D FT in Cartesian coordinate:
dudvevuFyxf yvxuj )(2),(),(
• Inverse 2D FT in Polar coordinate:
dudvevuyxf ),(),(
)sincos(2)i()( ddFf yxj
=l=G()
• Proof of filtered backprojection algorithm
20 0
)sincos(2)sin,cos(),( ddeFyxf yxj
• Proof of filtered backprojection algorithm
Inverse FT
EL5823 CT-1 Yao Wang, NYU-Poly 35
![Page 36: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/36.jpg)
Filtered Backprojection Algorithm• Algorithm:
– For each • Take 1D FT of g(l,) for each -> G()• Frequency domain filtering: G() -> Q(||G()• Take inverse 1D FT: Q() -> q(l) • Backprojecting q(lto image domain -> b(x,y)
– Sum of backprojected images for all
EL5823 CT-1 Yao Wang, NYU-Poly 36
![Page 37: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/37.jpg)
Function of the Ramp Filter• Filter response:
– c() =||– High pass filter
• G() is more densely sampled when is small andsampled when is small, and vice verse
• The ramp filter compensate for the sparser sampling at higher
EL5823 CT-1 Yao Wang, NYU-Poly 37
![Page 38: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/38.jpg)
Convolution Backprojection• The Filtered backprojection method requires taking 2 Fourier
transforms (forward and inverse) for each projectionI t d f f i filt i i th FT d i f l ti• Instead of performing filtering in the FT domain, perform convolution in the spatial domain
• Assuming c(l) is the spatial domain filter– || <-> c(l)– ||G() <-> c(l) * g(l)
• For each :– Convolve projection g(l) with c(l): q(l,)= g(l) * c(l)– Backprojecting q(lto image domain -> b(x,y)– Add b(x,y) to the backprojection sum
• Much faster if c(l) is short– Used in most commercial CT scanners
EL5823 CT-1 Yao Wang, NYU-Poly 38
![Page 39: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/39.jpg)
EL5823 CT-1 Yao Wang, NYU-Poly 39
![Page 40: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/40.jpg)
Step 1: Convolution
EL5823 CT-1 Yao Wang, NYU-Poly 40
![Page 41: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/41.jpg)
Step 2: Backprojection
EL5823 CT-1 Yao Wang, NYU-Poly 41
![Page 42: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/42.jpg)
Step 3: Summation
EL5823 CT-1 Yao Wang, NYU-Poly 42
![Page 43: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/43.jpg)
Ramp Filter Design
EL5823 CT-1 Yao Wang, NYU-Poly 43
![Page 44: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/44.jpg)
Practical Implementation• Projections g(l, are only measured at finite intervals
– ln; – chosen based on maximum frequency in G( ) W chosen based on maximum frequency in G(,), W
• 1>=2W or <=1/2W (Nyquist Sampling Theorem)• W can be estimated by the number of cycles/cm in the projection direction in the most detailed
area in the slice to be scanned
For filtered backprojection:• For filtered backprojection:– Fourier transform G(,) is obtained via FFT using samples g(n, )– If N sample are available in g, 2N point FFT is taken by zero padding g(n, )
• For convolution backprojectionFor convolution backprojection– The ramp-filter is sampled at ln– Sampled Ram-Lak Filter
2
evennoddnn
nnc
;0;/1
0;4/1)( 2
2
EL5823 CT-1 Yao Wang, NYU-Poly 44
![Page 45: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/45.jpg)
The Ram-Lak Filter (from [Kak&Slaney])
EL5823 CT-1 Yao Wang, NYU-Poly 45
![Page 46: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/46.jpg)
Common Filters• Ram-Lak: using the rectangular window• Shepp-Logan: using a sinc windowpp g g• Cosine: using a cosine window• Hamming: using a generalized Hamming window• See Fig. B.5 in A. Webb, Introduction to biomedical
imaging
EL5823 CT-1 Yao Wang, NYU-Poly 46
![Page 47: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/47.jpg)
Matlab Implementation • MATLAB (image toolbox) has several built-in functions:
– phantom: create phantom images of size NxNI PHANTOM(DEF N) DEF ‘Sh L ’ ’M difi d Sh L ’I = PHANTOM(DEF,N) DEF=‘Shepp-Logan’,’Modified Shepp-Logan’
Can also construct your own phantom, or use an arbitrary image
– radon: generate projection data from a phantom• Can specify sampling of p y p gR = RADON(I,THETA)
The number of samples per projection angle = sqrt(2) N
– iradon: reconstruct an image from measured projections• Uses the filtered backprojection method• Uses the filtered backprojection method• Can choose different filters and different interpolation methods for
performing backprojection[I,H]=IRADON(R,THETA,INTERPOLATION,FILTER,FREQUENCY_SCALING,OUTPUT_SIZE)
– Use ‘help radon’ etc. to learn the specifics– Other useful command:
• imshow imagesc colormap
EL5823 CT-1 Yao Wang, NYU-Poly 47
• imshow, imagesc, colormap
![Page 48: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/48.jpg)
Summary• Different generations of CT machines:
– Difference and pros and cons of each• X ray source and detector design• X-ray source and detector design
– Require (close-to) monogenic x-ray source• Relation between detector reading and absorption properties of the
imaged sliceimaged slice– Line integral of absorption coefficients (Radon transform)
• Reconstruction methods– Backprojection summationp j– Fourier method (projection slice theorem)– Filtered backprojection– Convolution backprojection
Equivalent, but differ in computation
• Impact of number of projection angles on reconstruction image quality
• Matlab implementations
EL5823 CT-1 Yao Wang, NYU-Poly 48
![Page 49: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/49.jpg)
Reference• Prince and Links, Medical Imaging Signals and Systems,
Chap 6.p• Webb, Introduction to biomedical imaging, Appendix B.• Kak and Slanley, Principles of Computerized
T hi I i IEEE P 1988 Ch 3Tomographic Imaging, IEEE Press, 1988. Chap. 3– Electronic copy available at http://www.slaney.org/pct/pct-
toc.html
EL5823 CT-1 Yao Wang, NYU-Poly 49
![Page 50: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/50.jpg)
Homework • Reading:
– Prince and Links, Medical Imaging Signals and Systems, Chap 6 S 6 1 6 3 36, Sec.6.1-6.3.3
• Note down all the corrections for Ch. 6 on your copy of the textbook based on the provided errata.
• Problems for Chap 6 of the text book:– P6.5– Consider a 4x4 image that contains a diagonal lineConsider a 4x4 image that contains a diagonal line
I=[0,0,0,1;0,0,1,0;0,1,0,0;1,0,0,0]; • a) determine its projections in the directions: 0, 45,90,135 degrees. • b) determine the backprojected image from each projection;• c) determine the reconstructed images by using projections in the 0
and 90 degrees only.• d) determine the reconstructed images by using all projections.
Comment on the difference from c)
EL5823 CT-1 Yao Wang, NYU-Poly 50
Comment on the difference from c).
![Page 51: Computed TomographyComputed Tomography (part I)cai2r.net/sites/default/files/documents/BIGP_Docs/...Convolution Backprojection • The Filtered backprojection method requires taking](https://reader033.fdocuments.us/reader033/viewer/2022060912/60a69a36ffeedc4ca763deda/html5/thumbnails/51.jpg)
Computer AssignmentDue: Two weeks from lecture date
1. Learn how do ‘phantom’.’radon’,’iradon’ work; summarize their functionalities. Type ‘demos’ on the command line, then select ‘toolbox -> image processing -> transform -> reconstructing an image from projection
‘ f fdata’. Alternatively, you can use ‘help’ for each particular function.2. Write a MATLAB program that 1) generate a phantom image (you can use
a standard phantom provided by MATLAB or construct your own), 2) produce projections in a specified number of angle, 3) reconstruct the p p j p g , )phantom using backprojection summation; Your program should allow the user to specify the number of projection angle. Run your program with different number of projections for the same view angle, and the different view angles, and compare the quality. You should NOT use the ‘radon( )’ g , p q y ( )and ‘iradon()’ function in MATLAB.
3. Repeat 1 but uses filter backprojection method for step 3). In addition to the number of projection angles, you should be able to specify the filter among several filters provided by Matlab and the interpolation filters usedamong several filters provided by Matlab and the interpolation filters used for backprojection. Compare the reconstructed image quality obtained with different filters and interpolation methods for the same view angle and number of projections. You can use the “iradon()” function in MATLAB
4 (Optional) Repeat 3 but uses convolution backprojection method You
EL5823 CT-1 Yao Wang, NYU-Poly 51
4. (Optional) Repeat 3 but uses convolution backprojection method. You have to do your own program.