[Gate-users] ARF fails to produce output

Tue Dec 3 15:50:29 CET 2019

Hi David,

So if I understand correctly, the ARF could be generated with a 2π (semi-isotropic) source that spans the area of the FOV in order to sample the angular response from all directions relevant to an image acquisition. Once the ARF is computed, I could place a source anywhere in front of the detector and the ARF will determine the response to a given photon based on its pos/dir/ene when impinging the detector plane.

When considering scattered photons, I would sample energy windows in intervals below the photopeak, but would it be best to use photons with a uniform energy distribution to sample the energy response from all directions? In that case, the same ARF table used as a scatter table for one isotope/simulation could be a photopeak table for a different isotope/simulation.

I agree, it is not the most convenient, and I am struggling to obtain any root output. The total number of primaries is updated while the Hits and Singles trees are empty.

I would be very grateful for your assistance with neural networks.

Thank you,
Matthew

________________________________
From: David Sarrut <David.Sarrut at creatis.insa-lyon.fr>
Sent: December 3, 2019 2:22 AM
To: Matthew Strugari <matthew.strugari at dal.ca>
Cc: gate-users at lists.opengatecollaboration.org <gate-users at lists.opengatecollaboration.org>
Subject: Re: [Gate-users] ARF fails to produce output

Hi Matthew,

1. not fully clear for me, but I think that the source should cover the whole range of a) detector position, b) direction, c) energy, that your detector will have to handle during the simulation that will use ARF. The idea is to build a statistical model (histogram) that is representative. My feeling is that the rectangular shape is not important, as soon as you cover the space (pos, dir, ene).

2. yes, once computed, any source can be used. The ARF is a statistical detector response model that gives you the probability to detect a given pos/dir/ene particle reaching the detector entrance plane in *all* of the energy windows of your detector. So it will work for any type of source, provided that the pos/dir/ene of photon impinging the detector plane is included in the range you learn in step 1.

We had some experiences with ARF and also struggled with the limitations (YZ plane etc). It works well but is not very convenient to use.

More recently, we have been working on a different version using neural networks. It is yet in development and requires dependences (libtorch) but if you are interested I will be happy can guide you.

David

On Mon, Dec 2, 2019 at 10:28 PM Matthew Strugari <matthew.strugari at dal.ca<mailto:matthew.strugari at dal.ca>> wrote:
Hi again everyone,

As it turns out, the ARF only works with the detector in the YZ plane. I was able to get the ARF working for my own application but the functionality is still not clear to me so any information on the following would be greatly appreciated.

  1.  Is the ARF data best generated with a rectangular source? According to "Implementation of angular response function modeling in SPECT simulations with GATE" by Descourt et al., the ARF tables are generated with a rectangular source but no rationale is provided.

  2.  Furthermore, once the ARF is tabulated and computed for a given collimator/detector pair and energy window, the ARF table may be used for any simulation involving the collimator/detector pair. Therefore, can the ARF tables be used in any simulation with any source definition? I would assume that the ARF tables should be generated under the same conditions for which they will be used (i.e. identical sources and detector placement in the generation and use of ARF tables) but this doesn't appear to be the case.

Regards,
Matthew

________________________________
From: Gate-users <gate-users-bounces at lists.opengatecollaboration.org<mailto:gate-users-bounces at lists.opengatecollaboration.org>> on behalf of Matthew Strugari <matthew.strugari at dal.ca<mailto:matthew.strugari at dal.ca>>
Sent: November 29, 2019 5:34 PM
To: gate-users at lists.opengatecollaboration.org<mailto:gate-users at lists.opengatecollaboration.org> <gate-users at lists.opengatecollaboration.org<mailto:gate-users at lists.opengatecollaboration.org>>
Subject: [Gate-users] ARF fails to produce output

Hi everyone,

Do any of you have a working example of the angular response functions from the ARF or SPECT_FFD examples in the GateContrib folders?

I could also use some clarification on a couple of things. First, the ARF example contains some errors/unclear settings such as the generation of the ARF tables using 364.5 keV photons and then the computation of the ARF tables using a reference energy of 140. keV. I assume this is a typo and 364.5 keV should be the reference energy but this discrepancy is present in both the readthedocs and the ARF macro files. Second, the readthedocs doesn't instruct users to set

  *   /gate/systems/SPECThead/setProjectionPlane <> <Unit>

where <> <Unit> "sets the YZ projection plane relative to the ARF device center". Is this line no longer required and does ARF only work with detectors in the YZ plane? My setup places my detector in the ZX plane and I cannot get any output to be produced. In all cases, the analog simulation produces output but the ARF simulation does not. I have attached the terminal output following execution of the SPECT_FFD angularResponseFunctionSPECT.mac simulation. It appears as though the ARF tables are not being loaded correctly.

Regards,

Matthew

--

Matthew Strugari

Biomedical Translational Imaging Centre - BIOTIC,
5890 University Ave,
Halifax, NS, B3K 6R8

_______________________________________________
Gate-users mailing list
Gate-users at lists.opengatecollaboration.org<mailto:Gate-users at lists.opengatecollaboration.org>
http://lists.opengatecollaboration.org/mailman/listinfo/gate-users

--
David Sarrut, Phd
Directeur de recherche CNRS
CREATIS, UMR CNRS 5220, Inserm U1206
Centre de lutte contre le cancer Léon Bérard
28 rue Laënnec, 69373 Lyon cedex 08
Tel : 04 78 78 51 51 / 06 74 72 05 42
http://www.creatis.insa-lyon.fr/~dsarrut
_________________________________
 "2 + 2 = 5,  for extremely large values of 2"
_________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/pipermail/gate-users/attachments/20191203/56bad728/attachment-0001.html>