Aims and objectives
Even if the image reconstruction of a computer tomography measurement is a simple linear mathematical problem nonlinear physical effects occur like beam hardening.
They appear as streak and cupping artifacts.
A simple possibility to treat cupping artifacts in computed tomography is empirical water precorrection [1,2].
In this work the correction quality in dependence on the geometry of the measured object is evaluated.
Methods and materials
The polychromatic raw data q should be transformed to equivalent monochromatic data p by a polynomial correction function (see Fig. 1).
Due to physical reasons the zeroth order can be neglected.
Because the inverse Radon transformation R-1 is a linear function,
the coefficients can be determined in the image domain.
Water phantoms of arbitrary size and transverse geometry can be used for...
At first the resulting correction vectors for simulated phantoms with different geometrical shapes but same cross sectional area (±0.9%) are compared:
cylinder (d=200 mm)
elliptical clinder (a=120 mm; b=83.33 mm)
box (a=178 mm)
For different geometries deviations in the resulting correction are determined.
The more the calibration phantom´s geometry is cylindrical the lower is the difference between the corrected and a simply linearly calculated reconstruction.
Also the size of the calibration water phantom is important.
objects with small diameters lead to a very small correction because of their low attenuation.
The calibrated attenuation range therefore is smaller and a linear extrapolation is...
and Willi A.
"Empirical cupping correction: a first-order raw data precorrection for cone-beam computed tomography." Medical physics 33.5 (2006): 1269-1274.
and Willi A.
"Empirical water precorrection for cone-beam computed tomography." Nuclear Science Symposium Conference Record,
and Dev P.
"Semiempirical model for...