Motion-compensated reconstruction of gated cardiac SPECT images using a deformable mesh model

  • Authors:

  • Thibault Marin;Miles N. Wernick;Yongyi Yang;Jovan G. Brankov

  • Affiliations:

  • lllinois Institute of Technology, Medical Imaging Research Center, Department of Electrical and Computer Engineering, Chicago, lllinois;lllinois Institute of Technology, Medical Imaging Research Center, Department of Electrical and Computer Engineering, Chicago, lllinois;lllinois Institute of Technology, Medical Imaging Research Center, Department of Electrical and Computer Engineering, Chicago, lllinois;lllinois Institute of Technology, Medical Imaging Research Center, Department of Electrical and Computer Engineering, Chicago, lllinois

  • Venue:
  • ISBI'10 Proceedings of the 2010 IEEE international conference on Biomedical imaging: from nano to Macro
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

We propose an algorithm for iterative, motion-compensated reconstruction of cardiac-gated SPECT. Dose limitations in SPECT lead to high level of noise in the projection data and further in the reconstructed images. Several reconstruction techniques have been reported to mitigate for the noise effects but they process each time frame individually and do not account for data temporal correlation. Advanced methods that allow for motion-compensated noise reduction use uniformly sampled pixels grid to represent images. Here we present a motion-compensated 4D reconstruction algorithm using content adaptive deformable mesh model (which is based on a deformable non-uniform sampling grid) for cardiac-gated SPECT. The proposed method tracks myocardial motion and utilizes the estimated motion to apply a cardiac-motion compensated temporal smoothing constraint during reconstruction. The temporal constraint is enforced between iterations of mesh based maximumlikelihood expectation-maximization algorithm. Specifically, temporal filtering is applied, in mesh domain, along the motion trajectory between iterations. The motion trajectory is estimation using our previously reported deformable mesh motion estimation technique. Visual comparisons as well as quantitative evaluation show that the proposed method achieves better noise reduction compared to several clinically used methods.