Affiliations: The CT laboratory, School of Mathematical Sciences,
Capital Normal University, Beijing, China | School of Software, Beijing Institute of Technology,
Beijing, China | Beijing Center for Mathematics and Information
Interdisciplinary Sciences, Beijing, China
Note: [] Corresponding author: Xing Zhao, The CT laboratory, School of
Mathematical Sciences, Capital Normal University, Beijing 100048, China.
E-mail: [email protected]
Abstract: Today's clinical dual energy computed tomography (DECT) scanners
generally measure different rays for different energy spectra and acquire
spatial mismatched raw data sets. The deficits in clinical DECT technologies
suggest that mainly image based material decomposition methods are in use
nowadays. However, the image based material decomposition is an approximate
technique, and beam hardening artifacts remain in decomposition results. A
recently developed image based iterative method for material decomposition from
inconsistent rays (MDIR) can achieve much better image quality than the
conventional image based methods. Inspired by the MDIR method, this paper
proposes an iterative method to indirectly perform raw data based DECT even
with completely mismatched raw data sets. The iterative process is initialized
by density images that were obtained from an image based material
decomposition. Then the density images are iteratively corrected by comparing
the estimated polychromatic projections and the measured polychromatic
projections. Only three iterations of the method are sufficient to greatly
improve the qualitative and quantitative information in material density
images. Compared with the MDIR method, the proposed method needs not to perform
additional water precorrection. The advantages of the method are verified with
numerical experiments from inconsistent noise free and noisy raw data.
Keywords: Dual energy computed tomography, iterative algorithm, basis material decomposition, calibration
