Time Of Flight reconstruction by NikEfth · Pull Request #304 · UCL/STIR

NikEfth · 2018-12-06T16:15:54Z

Introduces TOF reconstuction in STIR

KrisThielemans · 2019-01-18T18:09:55Z

@NikEfth, did you run the recon_test_pack locally? Travis throws up a lot of problems in it. For instance core dumps in the ROOT tests, but also others. it's always difficult to know with Travis of course.

KrisThielemans

a few house-keeping comments. nothing actually important.

KrisThielemans · 2018-12-06T16:35:24Z

src/include/stir/ProjDataInfo.h

  //! Get number of tof bins
  inline int get_tof_bin(const double& delta) const;
+
+  inline int get_unmashed_tof_bin(const double& delta) const;


don't use references for float, double etc. No point and it could slow it down.

add a doxygen comment

I don't think we should have this function. For some scanners this is not achievable.

KrisThielemans · 2019-01-19T11:18:53Z

recon_test_pack/run_test_time_of_fligh.sh

@@ -0,0 +1,106 @@
+#! /bin/sh


correct name of file

scripts/plot_TOF_bins.m

KrisThielemans · 2019-01-19T11:20:37Z

src/CMakeLists.txt

 option(BUILD_SHARED_LIBS 
       "Use shared libraries" OFF)

-### Settings for external libraries


somehow removed. re-instate I guess (I haven't checked properly)

This was weird. I merged with master and still this line did not re-appear. Only when I checked out the file.

KrisThielemans · 2019-01-19T11:23:13Z

src/IO/InputStreamFromROOTFileForCylindricalPET.cxx

    offset_dets = _offset_dets;

-    half_block = module_repeater_y * submodule_repeater_y * crystal_repeater_y / 2  - 1;
+     half_block = static_cast<int>( (module_repeater_y * submodule_repeater_y * crystal_repeater_y) / 2);


a bit weird to have a cast. The variables are all ints?

I was getting a warning. But seems to be ok.

KrisThielemans · 2019-01-19T11:24:50Z

src/IO/InputStreamFromROOTFileForCylindricalPET.cxx

        error("InputStreamFromROOTFileForCylindricalPET: The total number of entries in the ROOT file is zero. Abort.");

+    return Succeeded::yes;
+    half_block = static_cast<int>( (module_repeater_y * submodule_repeater_y * crystal_repeater_y) / 2);


code never executed

src/IO/InterfileHeader.cxx

KrisThielemans · 2019-01-19T12:30:46Z

src/IO/interfile.cxx

-	 for (seg++; seg != segment_sequence.end(); seg++)
-	   output_header << "," <<proj_data_info_ptr->get_max_ring_difference(*seg);
-	 output_header << "}\n";
+     std::vector<int>::const_iterator seg = segment_sequence.begin();


more white space

How would you like me to address this? Revert the file and reapply the changes?

KrisThielemans · 2019-01-19T12:35:10Z

src/buildblock/ML_norm.cxx

-              (*segment_ptr)[bin.axial_pos_num()][bin.view_num()][bin.tangential_pos_num()];
+	  for (bin.timing_pos_num() = proj_data.get_min_tof_pos_num(); bin.timing_pos_num() <= proj_data.get_max_tof_pos_num(); ++ bin.timing_pos_num())
+	  {
+		segment_ptr.reset(new SegmentBySinogram<float>(proj_data.get_segment_by_sinogram(bin.segment_num(),bin.timing_pos_num())));


this seems incorrect. fan-sums etc should sum over TOF bins. with current code. the fan data seems to be the one from the last TOF bin.

adaptations in ML_norm seems to need more thought. Can we just error out for now if its TOF?

Sure. Done.

NikEfth

I applied almost all comments.
The recon_test_past should not be able to function, yet.
One more commit for that.

NikEfth · 2019-01-29T04:09:21Z

recon_test_pack/run_test_time_of_fligh.sh

@@ -0,0 +1,106 @@
+#! /bin/sh


scripts/plot_TOF_bins.m

NikEfth · 2019-02-08T02:37:40Z

src/CMakeLists.txt

 option(BUILD_SHARED_LIBS 
       "Use shared libraries" OFF)

-### Settings for external libraries


This was weird. I merged with master and still this line did not re-appear. Only when I checked out the file.

NikEfth · 2019-02-08T02:42:42Z

src/IO/InputStreamFromROOTFileForCylindricalPET.cxx

    offset_dets = _offset_dets;

-    half_block = module_repeater_y * submodule_repeater_y * crystal_repeater_y / 2  - 1;
+     half_block = static_cast<int>( (module_repeater_y * submodule_repeater_y * crystal_repeater_y) / 2);


I was getting a warning. But seems to be ok.

src/IO/InputStreamFromROOTFileForCylindricalPET.cxx

src/IO/InterfileHeader.cxx

NikEfth · 2019-02-08T03:08:53Z

src/include/stir/ProjDataInfo.h

  //! Get number of tof bins
  inline int get_tof_bin(const double& delta) const;
+
+  inline int get_unmashed_tof_bin(const double& delta) const;


NikEfth · 2019-02-08T04:03:32Z

src/buildblock/ML_norm.cxx

-              (*segment_ptr)[bin.axial_pos_num()][bin.view_num()][bin.tangential_pos_num()];
+	  for (bin.timing_pos_num() = proj_data.get_min_tof_pos_num(); bin.timing_pos_num() <= proj_data.get_max_tof_pos_num(); ++ bin.timing_pos_num())
+	  {
+		segment_ptr.reset(new SegmentBySinogram<float>(proj_data.get_segment_by_sinogram(bin.segment_num(),bin.timing_pos_num())));


Sure. Done.

NikEfth · 2019-02-08T04:20:03Z

src/IO/interfile.cxx

-	 for (seg++; seg != segment_sequence.end(); seg++)
-	   output_header << "," <<proj_data_info_ptr->get_max_ring_difference(*seg);
-	 output_header << "}\n";
+     std::vector<int>::const_iterator seg = segment_sequence.begin();


How would you like me to address this? Revert the file and reapply the changes?

Fix a missing function declaration.

All tests pass. I have not checked the recon_tests yet

KrisThielemans · 2019-11-13T12:04:20Z

current Travis failures

The command "./run_test_listmode_recon.sh" exited with 1.
The command "./run_test_zoom_image.sh" exited with 1.
${PYTHON} -m pytest .
collecting 0 items / 1 errors/home/travis/.travis/functions: line 109: 18344 Segmentation fault

The run_test_zoom_image.sh is entirely weird as it doesn't have anything to do with TOF, so maybe ignore that here.

re-enabling some tests in test_proj_data_in_memory

KrisThielemans · 2020-01-29T21:38:59Z

run_test_listmode_recon.sh and the Python test are currently failing

KrisThielemans · 2020-02-03T18:31:56Z

@ALEXJAZZ008008 @NikEfth in my previous commit, I changed ProjData::get_num_sinograms() to include the num-TOF-bins. I did this because we call a Sinogram as 2D array. It also meant that size_all() could be implemented in terms of that, without having to know about TOF. However, this broke create_array_for_proj_data in swig/stir.i as you'll see.

It's easily fixable, but the question is what get_num_sinograms should return. I still think the name implies my change is the best, but feel free to disagree.

This did show that there are currently no Python (or of course MATLAB) tests for the TOF code. Maybe @ALEXJAZZ008008 can contribute some...

KrisThielemans · 2020-02-03T18:32:56Z

oh, maybe the Python test does check it, as I see my comment above that it's failing... Also, listmode recon...

NikEfth · 2020-02-04T13:04:05Z

I think get_num_sinograms should return the total number of 2D sinograms. That would include tof_poss and segments. Then individual functions will should be able to return the 3D or TOF positions.

However, now this feels a bit counter-intuitive. Because by default the tof positions are the "outer" layer of our ProjData. Not a big deal, though.

NikEfth · 2020-02-04T13:07:50Z

run_test_zoom_image fails, too

cannot test right now. Let's see what travis will say.

KrisThielemans · 2020-02-04T20:26:47Z

src/test/test_time_of_flight.cxx

 class cache_index{
 public:
-    cache_index() {
+    cache_index():


this doesn't seem to change anything at all, but best to use constructor initialisers. I don't see where cache_index is used by the way

I did this for codacy.

great. but is this class actually used?(sorry, can't see it)

KrisThielemans · 2020-02-04T20:45:01Z

src/swig/stir.i

-      int num_sinos = proj_data.get_num_sinograms();
-
-      Array<4,float> array(IndexRange4D(proj_data.get_num_tof_poss(),num_sinos, proj_data.get_num_views(), proj_data.get_num_tangential_poss()));
+      Array<4,float> array(IndexRange4D(proj_data.get_num_tof_poss(),proj_data.get_num_segments(), proj_data.get_num_views(), proj_data.get_num_tangential_poss()));


incorrect. This would have to be

const int num_non_tof_sinos = proj_data.get_num_sinograms() / proj_data.get_num_tof_poss();

and then use that. I'll fix this.

t.b.h. then I do not like this.
We have to make sure that under no situation the get_num_for_poss() does not return zero.
~~Which currently is the default value for the creation of a ProjDataInfo....~~
The mashing factor is zero.

I've done this slightly differently (by introducing get_num_non_tof_sinograms), but get_num_tof_poss() should never be zero of course. I guess we'd need a test for this then.

I don't understand what you mean with "The mashing factor is zero" (but possibly it's no longer relevant)

I took these values from #11, but cannot check them.

TOF clean-up

…_2024_01_02

…1_02 Merge master to TOF 2024 01 02

Add facility to add an alias for a keyword. This just works by looking-up the keyword in a map. This meansss type and callback remain the same. Also minor clean-up of test_KeyParser

use hopefully clearer names: - new: "TOF mashing factor", old: "%TOF mashing factor" - new: "Maximum number of (unmashed) TOF time bins", old: "Number of TOF time bins" - new: "Size of unmashed TOF time bins (ps)", old: "Size of timing bin (ps)" Old keywords are entered as aliases until STIR 7.0

changed TOF-related interfile keywords

Use get_bin_for_det_pos_pair instead. Addresses UCL#105

make get_bin_for_det_pair private and doc TOF argument

Fixes line width in python documentation, markdown bullet point spacing and Prettier formatting, and default nonTOF_distance_threshold in consistency test to 0.f

We were still writing "%TOF mashing factor".

also add changes to this function to release notes

was "Timing resolution (ps)"

also use new names for the keywords

this isn't currently used for parsing though

this is a work_around for UCL#1302

TOF interfile and scanner fixes

KrisThielemans

This will have to do for now. Outstanding issues and GE data will have to be handled on master.

Thanks to everyone!

KrisThielemans · 2024-01-02T08:50:09Z

src/include/stir/geometry/line_distances.h

+*/
+template <class coordT>
+inline void
+project_point_on_a_line(


@robbietuk could you check this? Maybe we can remove some duplication.

NikEfth requested a review from KrisThielemans December 6, 2018 16:15

KrisThielemans requested changes Jan 19, 2019

View reviewed changes

NikEfth commented Feb 8, 2019

View reviewed changes

NikEfth added enhancement in-progress labels Feb 8, 2019

KrisThielemans mentioned this pull request Feb 23, 2019

port STIR TOF support to SIRF SyneRBI/SIRF#315

Open

NikEfth requested a review from KrisThielemans March 3, 2019 04:31

NikEfth self-assigned this Mar 7, 2019

NikEfth and others added 7 commits April 11, 2019 17:10

Merge remote-tracking branch 'upstream/master' into tof_sino_UCL

d3aa6fc

Two more changes asked by Codacy

cb2c46c

Number of time frames and function declaration (#12)

a4b815e

Fix a missing function declaration.

Fix for TOF-PSF reconstruction

7c5ec7b

Merge remote-tracking branch 'upstream/master' into tof_sino_UCL

de8282b

Merge remote-tracking branch 'upstream/master' into tof_sino_UCL

3a25017

Merge with master

6d6b44b

All tests pass. I have not checked the recon_tests yet

KrisThielemans added 2 commits January 29, 2020 16:17

fix ProjData::get_num_sinograms() for TOF

2c068c2

Merge branch 'master' into tof_sino_UCL

a296248

re-enabling some tests in test_proj_data_in_memory

NikEfth added 4 commits February 4, 2020 14:37

fix sensitivity calculation for lm recon

8fa3169

Merge remote-tracking branch 'origin/tof_sino_UCL' into tof_sino_UCL

000a398

Fixed test_zoom_images

53085f0

fix: python

31dea95

cannot test right now. Let's see what travis will say.

KrisThielemans reviewed Feb 4, 2020

View reviewed changes

KrisThielemans and others added 12 commits January 2, 2024 09:26

replaced get_scanner_sptr() with get_scanner() for ListModeData

3ef019c

added Vereos but disabled for now as unchecked [ci skip]

db42d85

I took these values from #11, but cannot check them.

Merge pull request #59 from KrisThielemans/minor_TOF_fixes

d5aae0d

TOF clean-up

Merge remote-tracking branch 'origin/master' into merge_master_to_TOF…

c653a1b

…_2024_01_02

Merge pull request #61 from KrisThielemans/merge_master_to_TOF_2024_0…

d60ba7b

…1_02 Merge master to TOF 2024 01 02

tiny bit of TOF release notes [ci skip]

a82bd90

add KeyParser::add_alias_key

3e1561a

Add facility to add an alias for a keyword. This just works by looking-up the keyword in a map. This meansss type and callback remain the same. Also minor clean-up of test_KeyParser

update glossary for TOF and a few other things [ci skip]

22949cb

Merge pull request #62 from KrisThielemans/TOF_interfile_keywords

fde976a

changed TOF-related interfile keywords

make ProjDataInfo*NoArcCorr::get_bin_for_det_pair private

27a2516

Use get_bin_for_det_pos_pair instead. Addresses UCL#105

document that get_bin_for_det_pair takes mashed_tof_bin arg

572509a

KrisThielemans mentioned this pull request Jan 4, 2024

remove obsolete functions in ProjDataInfoCylindrical and ProjDataInfoCylindricalNoArcCorr #105

Open

13 tasks

KrisThielemans and others added 5 commits January 4, 2024 12:13

Merge pull request #63 from KrisThielemans/get_bin_for_det_pair_private

6f63046

make get_bin_for_det_pair private and doc TOF argument

remove defaults for some TOF-range to avoid unexpected problems

4a285ad

some TOF Codacy warning fixes for the GATE consistency test

53dbdc1

Fixes line width in python documentation, markdown bullet point spacing and Prettier formatting, and default nonTOF_distance_threshold in consistency test to 0.f

removed some test-scanners from Scanner

4088ec3

TOF Interfile header writing fix

ddd3e0c

We were still writing "%TOF mashing factor".

KrisThielemans changed the title ~~Time Of Flight reconstuction~~ Time Of Flight reconstruction Jan 5, 2024

KrisThielemans and others added 8 commits January 5, 2024 09:22

change defaults in ProjDataInfo::ask_parameters and remove span=0

a93870d

also add changes to this function to release notes

changed Interfile header keyword to "TOF timing resolution (ps)"

2cccf63

was "Timing resolution (ps)"

decrease precision in checking of TOF keywords

293d0f6

also use new names for the keywords

nicer formatting of Scanner::parameter_info

0818f69

add example TOF interfile header

be414b2

use STIR6.0 when output an Interfile header to denote TOF changes

d7330dc

this isn't currently used for parsing though

revert change and allow span=0 and use it in the scatter tests

c6b36ed

this is a work_around for UCL#1302

Merge pull request #66 from KrisThielemans/TOF_Interfile_scanner_fixes

0e11a8f

TOF interfile and scanner fixes

KrisThielemans reviewed Jan 5, 2024

View reviewed changes

KrisThielemans merged commit 9fd7457 into UCL:master Jan 5, 2024

KrisThielemans added this to the v6.0 milestone Feb 6, 2024

Conversation

NikEfth commented Dec 6, 2018

Uh oh!

KrisThielemans commented Jan 18, 2019

Uh oh!

KrisThielemans left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

NikEfth Feb 8, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

NikEfth left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

NikEfth Feb 8, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

KrisThielemans commented Nov 13, 2019

Uh oh!

KrisThielemans commented Jan 29, 2020

Uh oh!

KrisThielemans commented Feb 3, 2020

Uh oh!

KrisThielemans commented Feb 3, 2020

Uh oh!

NikEfth commented Feb 4, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

NikEfth commented Feb 4, 2020

Uh oh!

NikEfth Feb 8, 2019 •

edited

Loading

NikEfth left a comment •

edited

Loading

NikEfth Feb 8, 2019 •

edited

Loading

NikEfth commented Feb 4, 2020 •

edited

Loading

KrisThielemans Feb 4, 2020 •

edited

Loading

NikEfth Feb 5, 2020 •

edited

Loading