Motion Detection and Correction in Computed Tomography and Magnetic Resonance Imaging

1
Motion Detection and Correction in Computed
Tomography and Magnetic Resonance Imaging

• Zhengyan Sun
• Dr.Linney
• Saint Marys University

2
Outline

• Introduction
• Magnetic resonance (MR) imaging
• Computed tomography (CT) imaging
• Method for motion detection correction in MRI
• Method for motion detection correction in CTI
• Conclusion

3
Introduction

• Tomography
• Projection
• MR projection(F) --------------------imag
  e
• CT projection(S)------Frequency----------
  image
• Fourier Transform
• Motion problem

F-1 T
FT
F-1 T
4
Magnetic Resonance Imaging (MRI)

• Definition
• - Use a powerful magnet and radio waves to
• produce detailed images of the body.
• History
• - Developed in 1980s
• Risk
• - Pregnant women should avoid MRI scan

5
Computed Tomography Imaging (CT)

• Definition
• - A process in which a ring of detectors
  encircles
• an object and an x-ray source rotates
  about the
• object.
• History
• - Introduced by Hounsfield in 1973.
• Risk
• - Radiation

6
Method for Translational motion in MRI

• Gerchberg-Saxton algorithm 1
• start with G1(kx,ky) s(kx,ky)
• gn(x,y) F-1Gn(kx,ky)
• gn(x,y) gn(x,y)
  if (x,y) ? a
• 0
  otherwise
• Gn1(kx,ky) F gn(x,y)
  

7
Method for Rotational motion in MRI (a)

• For overlap data 2
• 1. Divide the views in to P different groups.
• 2. Insert each group in to a zeroed NN complex
  matrix.
• 3. Calculate the inverse 2-D Fourier transform
  of these matrices.
• 4. Using bilinear interpolation, back-rotate
  each of the p different
• images by the known rotation angle.
• 5. Calculate the 2-D FT of the p different
  images compute d .
• 6. Calculate the weights associated with each
  corrected k-space
• value.
• 7. Compute the weighted average values at each
  grid point.

8
Method for Rotational motion in MRI (b)

• For missing data 2
• 1. Noniterative algorithm
• 2. Iterative algorithm (POCS)
• - Effective and efficient to estimate
  large
• number of missing data
• - Not converge to a unique solution.

9
Method for Deformational
motion in MRI

• Noll et al. 3
• - motion model
• - use autofocus technique to estimate unique
• correction parameters for each point in
  the
• image.

10
Method for Translational motion in CT

• Opposite Ray algorithm 4
• 1. Find opposite ray of each ray in sinogram
  domain.
• 2. Compute the difference of opposite rays and
• construct the motion.
• 3. Using the filter to reduce spatial errors.
• 4. The filtered sinogram of motion is
  integrated.
• 5. The integrated sinogram of motion is
  reconstructed.

11
Method for Rotational motion in CT

• Maximum likelihood adaptive neural system
• (MLANS) 5
• - The object function is modeled as a
  mixture of
• localized basis components
• - MLANS provides an iterative framework for
• optimize the model parameters

12
Method for Deformational motion in CT

• CTX algorithm 6
• Steps
• 1.find a time-varying motion model
• 2. reconstruct image by filtered
  back-projection
• Limitation not fit for the motion in chest
• Pixel-Specific Back-Projection algorithm
• 1.CTX model parameter are estimated for
  each point.
• 2. Reconstruction

13
Conclusion

• The talk involved
• MR and CT imaging
• Translational, rotational, deformational
  motion
• My research area

14
Reference

• 1 R.W.Gerchberg, W.O.Saxton, A practical
  algorithm for the determination of phase from
  image and diffraction plane pictures, Optik,
  vol.35, no.2, pp237-248,1972.
• 2 C.Weerasinghe, H.Yan, An improved algorithm
  for rotational motion artifact suppression in
  MRI, IEEE Trans. On Med.Imag.vol.17,no.2,1998
• 3 D.C.Noll, J.M.Pauly, C.H.Meyer, Deblurring
  for non-2-D Fourier transform magnetic resonance
  imaging, Magnet.Res.Med., vol25, pp.319-333,1992
• 4 N.C.Linney, P.H.Gergson, Organ motion
  detection in CT images using opposite rays in
  fan-beam projection systems, IEEE Trans.on
  Med.Imag. Vol20, no.11, 2001
• 5 Deming, R.W., Reconstruction of time-varying
  objects in computed tomography using a
  model-based neural network, Proceeding of the
  1998 IEEE International Symposium on,1998
• 6 C.R.Crawford, K.F.King, C.J.Ritchie,
  Respiratory compensation in projection imaging
  using a magnification and displacement model,
  IEEE Trans.on.Med.Imag. Vol15,no.3, 1996