FAQ

Looking for more information? The following information may be helpful for you.

• How can the users create the material-specific volume images (a_k) similar to the ones used for CT workflow (Fig. 9) of the PcTK paper (Ref. 1)? (03-19-2018, 04-17-2018)
  • Our procedure is described in Sec. 3.C of Ref. 1. Users are encouraged to develop better procedures and share their programs for the community spirit with us. Please send us an email if you are willing.
  • The material-specific volume images (a_xyz,k) for the “basis material” k we used for Ref. 1 was calculated as follows. The local linear attenuation coefficient at xyz, µ_xyz, are modeled by a linear combination of multiple basis materials k as µ_xyz = ∑_k r_xyz w_xyz,k (µ_k /r_k) = ∑ _k a_xyz,k µ_k, where r_xyz is the local density of the object at xyz, w_xyz,k is the local fraction-by-weight of the density of the material k, (µ_k /r_k) is the energy-dependent mass attenuation coefficient of the material k, and r_k is the density of the material k at the standard state. A material-specific line integrals v_k,j for material k and a sinogram pixel j can then be calculated by forward projecting a_xyz,k as v_k,j= ∑_xyz h_xyz,j a_xyz,k, where h_xyz,j is an element of the projection system matrix H_xyz and a product of v_k,j and µ_k is used to calculate the attenuated spectrum S_t,j for the sinogram pixel j. Some of the data can be found in the folder 1_inputdata or 5_refdata. But again, the above was merely our choice made to demonstrate PcTK and CT workflow; users should use their way to define a_xyz,k and calculate v_k,j.
  • The mass attenuation coefficients and other properties of various materials are provided in ICRU Report 44 as well as National Institute of Standards and Technology (NIST). 
• Can the number of material-specific sinograms for CT workflow more than 2? (03-22-2018)
  • Yes, it can be any large number or as few as 1. You just need to modify the workflow code to read data and use them to calculate the attenuated spectrum incident onto the PCD pixels, S_t,j(E₁). You also need to prepare (1) material-specific sinograms (data (C) in Sec. 2.C of PcTK paper), and (2) the corresponding energy-dependent linear attenuation coefficients for each material (data (D)).
  • It is advised to check the image quality of CT images at different energies such as 40, 70, 100, and 130 keV, prior to proceed and generate material-specific sinograms. Often artifacts, thus, problems, are more apparent with the material-specific density images (a_ST and a_BN in Fig. 9 of PcTK paper) than the material-specific sinograms.
• What are the geometry parameters used for CT workflow (Fig. 9) of the PcTK paper (Ref. 1)? (03-19-2018)
  • Focus-to-detector distance: 1,080 mm
  • Focus-to-iso-center distance: 600 mm
  • Detector pixel size at the detector (not the projected size at the iso-center): 225 µm
  • Detector shape: Cylindrical detector with equi-angular sampling
  • The other parameters (i.e., the number of detector channels and rows) must be apparent or can be calculated from the size of data (.mat files) or specified in Ref. 1.
• What is the format of .flt data? (03-19-2018)
  • It is 32-bit floating data with no header.
• Which forward projector would you recommend? (03-19-2018)
  • Any projector will be fine and please send us information on available projectors.
  • We used an in-house projector for Ref. 1 and are using ASTRA toolbox for other projects.