Other info related to STIR
Charalampos Tsoumpas (King's College London) has made a set of simulated 4D PET-MR datasets
available that can be used freely by anyone (academia and industry) for any scientific investigation.
You can find further information
on the relevant website.
STIR uses a modified version of Interfile for its native data format. The modifications
were intended to become part of Interfile 4.0 but it is not clear if this will ever
happen. See the original proposal as PDF
for more information on the modifications. (Note that currently, STIR does not support
images that are in strict Interfile version 3.3 format. Moreover, it only supports a subset
of the suggested new keywords).
GATE_PET_2_STIR is a project
(independent from STIR) to convert GATE data to STIR sinograms. Let us know how it works via the
The PARAPET project (precursor of STIR) has a number of deliverables that are still somewhat useful
for STIR users. You can find them here.
- PARAPET Deliverable 4.1: PARAPET buiding blocks documentation as PDF. Note: written in 1999.
Pretty much completely out of date with respect to STIR, but contains
slightly more documentation on the projectors.
Deliverable 1.3: PET reconstruction algorithm survey as PDF. Note: written in 1997.
- Deliverable 5.2 on OSMAPOSL has essentially the same content as
Jacobson M, Levkovitz R, Ben-Tal A, Thielemans K, Spinks T, Belluzzo D, Pagani E, Bettinardi V, Gilardi M C, Zverovich A and Mitra G Enhanced 3D PET OSEM Reconstruction using inter-update Metz filters, Phys. Med. Biol. 45 No.8 (2000) 2417-2439.
- Deliverable 5.4: Specification of Bayesian
Reconstruction Building Blocks as PDF. Note: written in 2000.
Deliverable 5.6a: Evaluation of iterative reconstructions as PDF. Note: written in 2000.
Deliverable 5.6b: Evaluation of analytic reconstructions as PDF. Note: written in 2000.
There are a number of other software packages available for image
reconstruction. This is an (incomplete?) overview with some subjective comments.
All these packages concentrate on image reconstruction for static imaging.
STIR is unique in providing a flexible open source framework for PET including scatter estimation and routines for dynamic imaging.
You could read the following review by Roberto de la Prieta
Free Software for PET Imaging
- Fessler and his students have implemented an
Image Reconstruction Toolbox
as a set of open source matlab routines for image reconstruction and processing. In addition, Fessler has also developed ASPIRE. However, this is distributed via executable files and its license allows non-commercial usage only.
provides an implementation of OSEM for PET image reconstruction using a well-designed GUI. It could also use reconstruction algorithms available in STIR 1.4 but it is not compatible with STIR 2 or later.
TIRIUS is distributed under the GNU Lesser General Public License.
- NiftyRec provides
a matlab (and also some python) interface to C/GPU projectors and some reconstruction
algorithms. Current projectors are mostly suitable for SPECT. License is BSD-style.
- EMrecon is another recently released package, using expectation maximisation. License is unclear
- QSPECT is a recently released package for SPECT reconstruction using MLEM or OSEM.
OSCaR (Rezvani et al, 2007) provides a matlab-based implementation of the FDK algorithm for Cone-beam CT and it is freely available to AAPM members, but it is not clear if non-members can also use it.
POSSUM is an MRI
simulator which also contains FBP.
CTsim is an analytic CT simulator (2D) which also contains FBP. It is not clear if this is still maintained as the
web-site does contain a lot of dead links.
OpenRTK is a recent reconstruction
package for (cone-beam) CT, based on ITK. License is Apache 2.0.
high-performance GPU primitives for 2D and 3D tomography. It supports 2D parallel and fan beam geometries, and 3D parallel and cone beam. All of them have highly flexible source/detector positioning.
A large number of 2D and 3D algorithms are available, including FBP, SIRT, SART, CGLS. License is GPL3.
- Troll: a
pyCuda based toolkit for PET, later (now?) SPECT and CT. In early development stage (2013). License is GPL3.
- regularization techniques for iterative reconstruction for matlab (using mex) developed as part of the Collaborative Computational Project in Tomographic Imaging which concentrates on CT. Their reconstruction software is planned to become public in 2015 or so.
- TomoPy: Python-based reconstruction software from
the ANL. License is BSD-style.
This page is licensed with the
Creative Commons Attribution-ShareAlike 3.0 Unported License
Last modified: Wed Apr 23 21:52:57 BST 2014