Based on a segmented MRI dataset of a normal subject (100 structures) including the ability to simulate rigid motion of the brain (3 translation, 3 rotation).
(Re)compile using make
, or use CMake
(e.g. for Visual Studio
).
Then run dncat_bin
(without arguments) for help:
$ make
$ ./dncat_bin
Program to produce voxelised phantoms from NURBS torso file input
Usage:
./dncat_bin [options] <gen_par> <out_base>
Options:
-x X translation in mm
-y Y translation in mm
-z Z translation in mm
-X X rotation in degrees
-Y Y rotation in degrees
-Z Z rotation in degrees
Arguments:
<gen_par> : general parameter filename. Note that start_slice should be
an even number > 1 to avoid artefacts.
<out_base> : base string for output files
./dncat_bin test.par resulting_image1
will create a 128x128x128
sample brain image (see
test.par
for definitions). The values we have used in this particular example are
somewhat close to a typical anatomic MR study (but this may need to be
better tuned and double-checked for the various regions).
- Paul Segars 2006-09-03
- Casper da Costa-Luis (casperdcl) 2017-05
Original can be downloaded from here.
Please reference the following work (in which the construction of the phantom has been more elaborately mentioned):
- Rahmim, K. Dinelle, J. C. Cheng, M. A. Shilov, W. P. Segars, S. C. Lidstone, S. Blinder, O. G. Rousset, H. Vajihollahi, B. M. W. Tsui, D. F. Wong, and V. Sossi, "Accurate event-driven motion compensation in high-resolution PET incorporating scattered and random events", IEEE Trans. Med. Imag., vol. 27, pp. 1018-1033, 2008.