CT Statistical and Iterative Reconstructions and Post Processing
For decades, CT images were reconstructed from the measured raw data using analytical reconstruction methods, such as filtered backprojection (FBP). While FBP is fast and accurate, the measured data are rarely ideal, and iterative and statistical methods can provide significant benefit particularly when the data quality is poor. They can produce images with lower noise and can also reduce artifacts from system imperfections. However, the resulting images are nonlinear and non-stationary and have other properties that are different from those produced by FBP that should be kept in mind when the images are interpreted.
KeywordsImage reconstruction Filtered backprojection Iterative reconstruction Statistical reconstruction Edge-preserving smoothing
- 3.Kak AC, Slaney M. Principles of computerized tomographic imaging. New York: IEEE Press; 1988.Google Scholar
- 8.Hsieh SS, Pelc NJ. Limits to dose reduction from iterative reconstruction and the effect of through-slice blurring. Proc SPIE Med Imaging. 2016; 97831C.Google Scholar
- 14.Olcott EW, Shin LK, Sommer G, Chan I, Rosenberg J, Molvin FL, et al. Model-based iterative reconstruction compared to adaptive statistical iterative reconstruction and filtered back-projection in CT of the kidneys and the adjacent retroperitoneum. Acad Radiol. 2014;21(6):774–84.CrossRefGoogle Scholar
- 17.Akagi M, Nakamura Y, Higaki T, Narita K, Honda Y, Zhou J, et al. Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT. Eur Radiol. 2019 (published online).Google Scholar
- 18.FDA Clears GE’s Deep Learning Image Reconstruction Engine. Imaging Technology News. 2019. Retrieved May 24, 2019, from https://www.itnonline.com/content/fda-clears-ges-deep-learning-image-reconstruction-engine.