GAMOS 5.1.0 User's Guide | ||
---|---|---|
<<< Previous | Image reconstruction utilities | Next >>> |
This program applies a SSRB-FBP2D 22 method to reconstruct PET images. First, a Single Slice rebinning algorithm (SSRB converts 3-dimensional projection data in a set of 2-dimensional sinograms, one per each axial image slice; next, a Filtered Back Projection (image reconstruction algorithm that makes use of the Radon transform 23 ) is performed in each slice.
As input data, the program requires the name of the GAMOS projection data file without extension (.hv/.v). These files can be generated by GAMOS (see "PET output for reconstruction" section) or converted from list-mode files using the lm2pd utility. The command-line options are the following:
-m: SSRB Maximum ring difference (by default: n_planes-1). The user can discard too axially tilted lines of response by selecting a lower value .
-r: SSRB Normalization (by default: 1=yes, 0=no).
-i: Image size, number of X and Y pixels (by default: number of sinogram bins).
-x: Transaxial pixel size (by default: sinogram bin size).
-f: Type of apodization window for filtering 24 (low frequency pass):
1, Generalized Hamming (value by default, it includes Hann window and ramplak).
2, Butterworth.
3, Shepp-Logan.
-a: Alpha parameter for generalized Hamming window (by default: 1, ramplak filter).
-b: Order for Butterworth window (by default 4).
-c: Cutoff frequency, relative to Nyquist freq: bin_size/2 (by default: 0.75).
-n: Name of the output image file ("MY_NAME", without extension). The program will write a "MY_NAME.img" data file and two associated header text files: "MY_NAME.hv" and "MY_NAME.hdr". The .hv file has a STIR-like Interfile format 27 , whereas the .hdr file uses Interfile 3.3 conventions; it is done to make possible opening the image by different programs (see "Visualization tools" section).
-h: Help, printing arguments list.
-v: Verbosity (by default:0 for silent, 3 for debugging).
The following example will apply a SSRB-FBP2D reconstruction to a projection data file named "MY_PROJDATA_ALL.hv" using a Butterworth filter of order=6 and cutoff frequency=0.5 Nyquist units, image and pixel sizes take values from the projection data (tangential bins):
ssrb_fbp MY_PROJDATA_ALL -f 2 -b 6 -c 0.5 -n MY_IMAGE |
To use this utility, you have to install the FFTW library at the path where the variable FFTW_BASE_DIR points (type echo $FFTW_BASE_DIR. You can download this library at http://www.fftw.org
<<< Previous | Home | Next >>> |
Summing projection data files: sumProjdata | Up | Visualization tools |