STIR Download Area

---

Contents

• Download current release
• Preliminary files
• Other pages of direct relevance
• How to report problems
• How to contribute
• Notes

Download current stable release

conda is the easiest way to install STIR. Check our Wiki page for help.

To get the source code and build STIR yourself, we highly recommend to use git:

    git clone --recurse-submodules https://github.com/UCL/STIR
    # this will put you on the master branch. If you want to switch to a release:
    git checkout --recurse-submodules rel_6.3.0
  

If you do not want to use git, please check STIR releases on GitHub. Scroll down to the bottom of a release for links for the source code.
Warning: The .tar.gz/.zip files on GitHub releases currently do not contain submodules such as fmt, needed for older compilers. Therefore, use git if you can.
See the User's Guide for more information on dependencies.

Below are links to documentation. You will probably want to shift-click, right-click, control-click or whatever on these, such that your browser does not attempt to open the archives.

• User's Guide as .pdf file (includes installation instructions), Read it!
• zipped software documentation (~20 MB)

Back to contents

Preliminary files

This section contains extra downloadable files that might be useful to developers. Some might come under special license restrictions. Use at your own risk, and check the mailing lists!

• Preliminary code for reading the LMF listmode format. This code is however incomplete and does not compile and needs some work.
• a zip file containing the original PARAPET code for parallellisation. This will not work with STIR in its current version unfortunately. Extract (NOT in your STIR subdirectory) and read READMEpara.txt.
  However, since version 2.1 STIR contains its own parallel code (using MPI) using similar concepts to the PARAPET code, so these files are obsolete.

Back to contents

Other pages of direct relevance

• Software documentation
• STIR Wiki
• AMIDE is a freely available image viewer and analysis program that can read the .hv files produced by STIR. Unfortunately, this no longer seems to work on recent Ubuntu.

Warning:
Because of some unfortunate incompatibilities between Interfile 3.3 and the version of Interfile used by STIR, both AMIDE and xmedcon ignore the file offset field in .hv files, so if your the binary data does not start at file offset zero, import into these programs will fail. You can then try the .ahv files, but this will ignore any scale factor in the header.
Luckily, all .hv files produced by STIR have file offset 0, except those produced by ifheaders_for_ecat7 or when reading DICOM or Nifti files. Of course, AMIDE and xmedcon read ECAT7 files directly.

Back to contents

How to report problems

The general rule is: the more effort you do yourself, the more chance you have that someone will help you to fix it.

1. Make sure it is not a known problem: check this page for patches, read the STIR Wiki and search the mailing lists. (Use your favorite search engine or the link on that page.)
2. Make sure it is not a problem with your data. For images, use list_image_info, manip_image. For projection data
   1. use list_projdata_info to see if you and STIR agree on what the data sizes etc are.
   2. use display_projdata (or manip_projdata) to look at all segments (not just a few).
3. In other cases, in particular if you have a segmentation fault, you will have to build the debug version of STIR and debug the problematic program to at least see where it crashes. For instance when using gdb, this goes as follows:

   gdb /wherever/it/is/STIR/debug/analytic/FBP2D/FBP2D 
   gdb> cd /wherever/the/data/is
   gdb> r myparameterfile
   (wait till segmentation fault)
   gdb> info stack
   (some more gdb commands)
   gdb> quit
   

   Ideally you then have a look at the value of some of the variables (use "p variablename"), maybe after moving up the call-stack ("up").
   Follows the next link for the GDB documentation.
4. Report the problem on the mailing lists. Please provide as much info as you can. In particular, also give details about OS and compiler version.

Back to contents

How to contribute

You are very (very) welcome to contribute to STIR. As STIR is Open Source, you can distribute your work (including the rest of STIR) yourself. However, it would of course be better to get your work included into the "official" STIR distribution from this web-site. See our page with more information.

Back to contents

Notes

Please register on our mailing lists. Questions about our software will ONLY be answered when directed to one of the mailing lists or via GitHub issues.

After extracting the files, please check STIR/buildblock/Scanner.cxx to see if the parameters of your scanner are correct. If not, let us know. You might want to tell us about some more names for your scanner. (These names are used to recognise the scanner in the Interfile header file (*.hs)).

Back to contents

---

Last modified: Jan 13 2025