They are all stored in float format one next to each other. Each n files has 232404 bytes (58101 floats) that correspond to the folowing fields:
geometric effects: 344x127 matrix. Changes of sensitivity due to geometric effects for each 2d sinograms. 344 is the s axes (sampling distance) and the 127 corresponds to projections. Each column are equals. So the only effect here depends on s. It needs to be resized to fit the sinogram size.
crystal interference: 9×344 matrix. Is the pattern generated because of the crystal interference in a block. Depends on the incident angle of the photon. For each crystal element in a block in a ring, there is a pattern. So this matrix has 9 rows (one for each crystal element and one additional for the gap between blocks) and 344 cols that are the amount of s samples of a sinogram. To use it, the matrix must be repeated in the rows direction to complete all the blocks (and have a omplete sinogram of 252x344).
crystal efficiencies: 504×64 matrix. This is the efficency of each crystal. 504 are the crystals inside a ring: (8 pixel element + 1 gap)*56 blocks. And the cols correspond to each ring.
axial effects: 837. One factor for each sinogram of the span11 sinograms. In the same order that they are stored in the interfile format. It can be observed that is related with the axial compression of span 11 sinograms (and maybe anothr effect)
paralyzing ring DT parameters: 64. Paralyzing dead time for each ring. At the moment is the same value for each ring. So it's constant.
non-paralyzing ring DT parameters: 64. Non-paralyzing dead time for each ring. At the moment is the same value for each ring. So it's constant.
TX crystal DT parameter: 9. Dead time of each crystal in a block in a ring. All values in zero.
additional axial effects: 837. Additional axial effects. One factor for each sinogram in the span 11 sinograms.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
First we will deal with the normalzation of the transverse sinograms.
The geometric component matrix can be seen in:
The crystal interference pattern and crystal efficencies are shown in:
With these data, a 2d sinogram is generated for the normalization correction factors (NCF) of the geometric effect and the crystal interference, and a sinogram including both (multiplication of each factor):
NCF for the transversal geometric component
NCF for the crystal interference component
NCF for the crystal interference and geometric components
For the crystal efficencies, a sinogram is generated from the 508 efficencies available for each ring. This need to be verified, the position of each crystal was taken from the sinogram viewer from siemens. Example of one sinogram with the NCF for crystal efficencies:
Example of NF for the crystal efficency component
The variation in the crystal efficencies is not big. In this plot the crystal efficencies for the 508 crystal elements of the first 10 rings are shown:
The .n file for normalization has 8 componentes:
They are all stored in float format one next to each other. Each n files has 232404 bytes (58101 floats) that correspond to the folowing fields:
geometric effects: 344x127 matrix. Changes of sensitivity due to geometric effects for each 2d sinograms. 344 is the s axes (sampling distance) and the 127 corresponds to projections. Each column are equals. So the only effect here depends on s. It needs to be resized to fit the sinogram size.
crystal interference: 9×344 matrix. Is the pattern generated because of the crystal interference in a block. Depends on the incident angle of the photon. For each crystal element in a block in a ring, there is a pattern. So this matrix has 9 rows (one for each crystal element and one additional for the gap between blocks) and 344 cols that are the amount of s samples of a sinogram. To use it, the matrix must be repeated in the rows direction to complete all the blocks (and have a omplete sinogram of 252x344).
crystal efficiencies: 504×64 matrix. This is the efficency of each crystal. 504 are the crystals inside a ring: (8 pixel element + 1 gap)*56 blocks. And the cols correspond to each ring.
axial effects: 837. One factor for each sinogram of the span11 sinograms. In the same order that they are stored in the interfile format. It can be observed that is related with the axial compression of span 11 sinograms (and maybe anothr effect)
paralyzing ring DT parameters: 64. Paralyzing dead time for each ring. At the moment is the same value for each ring. So it's constant.
non-paralyzing ring DT parameters: 64. Non-paralyzing dead time for each ring. At the moment is the same value for each ring. So it's constant.
TX crystal DT parameter: 9. Dead time of each crystal in a block in a ring. All values in zero.
additional axial effects: 837. Additional axial effects. One factor for each sinogram in the span 11 sinograms.
First we will deal with the normalzation of the transverse sinograms.
The geometric component matrix can be seen in:

The crystal interference pattern and crystal efficencies are shown in:

With these data, a 2d sinogram is generated for the normalization correction factors (NCF) of the geometric effect and the crystal interference, and a sinogram including both (multiplication of each factor):

NCF for the transversal geometric component
NCF for the crystal interference component
NCF for the crystal interference and geometric components
For the crystal efficencies, a sinogram is generated from the 508 efficencies available for each ring. This need to be verified, the position of each crystal was taken from the sinogram viewer from siemens. Example of one sinogram with the NCF for crystal efficencies:

Example of NF for the crystal efficency component
The variation in the crystal efficencies is not big. In this plot the crystal efficencies for the 508 crystal elements of the first 10 rings are shown:

Last edit: GPGPU-UTN-FRBA 2015-02-04