[Gate-users] Further documentation of output images created by the FixedForcedDetection Actor

Thu Sep 8 09:08:05 CEST 2016

Hi,
Sorry for the late reply. I'm not sure I understood your workaround... You
can put anything you'd like in one voxelized volume but my feeling is that
it is going to be difficult to define a voxelized volume containing the
filters since they are generally far apart, that would make huge voxelized
volume. If you meant severat voxelized volume daughter of one volume, that
will not work.
Please stay on the mailing list,
Simon

On Mon, Sep 5, 2016 at 9:52 AM, Lukas Gromann <lukas.gromann at tum.de> wrote:

> Hi Simon,
>
> Ohh, yes I missed that part in the Wiki, as it looked liked as „
> GateHybridForcedDetectionActor“, would be a new actor due to the name and
> separator in the Wiki.
>
> This makes everything a little more complex…
>
> Would this kind of workaround work?
> 1. I define a mother „CT Volume“ out of Air, and add the Patient-Phantom
> and e.g. the Patien-Bed and my Filters all as Offspring-Volumes of the
> „CT-Volume“.
> 2. Then I would do the Simulations to get the intensities for Primary,
> secondary etc.
> 3. Repeat the Simulations, with just the Patient-Phantom Subvolume
> disabled.
> 4. Calculate the Attenuation Images not from the given Flatfield, but
> rather take the Primary image from (3.) as my new Flatfield.
>
> Would this work? Or would this still violate the precondition, that there
> is more than one „voxelized“ volume?
>
> Thanks for your Help.
>
> Lukas
>
>
> ----------------------------------------------------
> Lukas Gromann
>
> Lehrstuhl für Biomedizinische Physik (E17)
> Technische Universität München
> Boltzmannstrasse 11
> 85748 Garching
>
> Lukas.Gromann at TUM.de
>
> Tel.: 089-289/10905
> Mobil: 017629718306
>
> Am 05.09.2016 um 09:30 schrieb Simon Rit <simon.rit at creatis.insa-lyon.fr>:
>
> Hi,
>
> Here come the limitations of the actor which is documented on the wiki.
> Currently, you can only have one voxelized volume in your scene, nothing
> more. So no filter. When we need to add a filter we change the spectrum in
> our use of the actor but, for example, we cannot add a bow tie filter. So
> the only thing that is removed from the beam path is this voxelized volume.
>
> Simon
>
> On 04/09/2016 22:25, Lukas Gromann wrote:
>
> Hi Simon,
>
> Thanks you for your explanations. This makes everything much clearer.
>
> A further question regarding the flatfield, which I consider a very useful
> feature: Is there a way to configure, which Volumes are kept in the
> beampath, when the flatfield intensities are calculated? I guess currently
> only the source and the detector are kept when calculating the flatfield?
> I would like to keep different type of filters and collimators in the FOV
> during the Flatfield „aquisition“.
> If there is no config parameter to control this, I guess a simple way
> would be a recalculation with just the phantom volume disabled? (At the
> price of doubling the calculation time)
>
> Thanks,
>
> Lukas
>
> ----------------------------------------------------
> Lukas Gromann
>
> Lehrstuhl für Biomedizinische Physik (E17)
> Technische Universität München
> Boltzmannstrasse 11
> 85748 Garching
>
> Lukas.Gromann at TUM.de
>
> Tel.: 089-289/10905
> Mobil: 017629718306
>
> Am 02.09.2016 um 12:24 schrieb Simon Rit <simon.rit at creatis.insa-lyon.fr>:
>
>
> On 02/09/2016 11:53, Lukas Gromann wrote:
>
> Hi Simon,
>
> thanks for your fast reply, this helps me already a lot. Indeed I read
> those parts in the wiki and had also a short look at the papers mentioned.
>
> Especially the following questions still arise for me:
> - What are the „signals“ in primary, total, compton etc? Are these
> detection probabilities? Photon counts? Or total recorded Photon-Energies
> per pixel, or is this converted to some arbitrary ADU-values?
>
> There is a detector response curve that will set this. This is a text file
> in which the first column is the incoming photon energy (in MeV) and the
> second column is the detected signal. It's used both in the Monte Carlo and
> the fixed forced detection examples, I first describe it for Monte Carlo
> where you have one photon arriving on the pixel at a time. If you don't set
> it, it will integrate the energy of the incoming photons, which would be
> similar to setting it to a ramp:
> 0 0
> 1 1
> Note that the values are linearly interpolated.
>
> The other option is to do a photon counting, so to add 1 regardless of the
> energy:
> 0 1
> 1 1
>
> You can also set a much more complex (but more realistic) function, as
> described in [Roberts et al, PMB, 2008] which we recently used in [Vilches-Freixas
> et al, Med Phys, 2016] <http://dx.doi.org/10.1118/1.4961400>. I have
> enclosed an example.
>
> Now, for fixed forced detection, the final value is also multiplied by the
> probability of having a photon with that energy in this pixel since this is
> what FFD computes.
>
> - I guess this is related to the energy-response of the detector?  Also
> here the example provides some basic responseDetector.txt, but the Syntax
> is not clear to me.
> - In the example Code, there is an option called
> „/gate/actor/ffda/energyResolvedBinSize        0 keV“, which is not
> mentioned in the wiki at all. If I would set this Value to 1keV, would I
> create a Photon counting detector with 120 energy bins? How would the
> Output look like? 120 images each for „compton“ and "primary“?
>
> Yes, these are recent (unreleased) dev. If you set this to a non 0 value,
> instead of storing 2D projection images, it stores 3D projection images
> where the third dimension is the energy. There is still one projection only
> per gantry angle. Indeed, if your max spectrum energy, it would create 120
> energy bins with a perfect energy resolved detector.
>
>
>
>
> I would therefore suggest to add at least the following lines to the
> documentation:
>
> All contributions are summed in total which can be decomposed in
> primary+secondary. Secondary (same as scatter) can itself be decomposed in
> compton+rayleigh+fluorescence. flatfield is available to compute the
> measured primary signal if there is no object, this is useful for CT to
> apply the beer Lambert law. attenuation is ln(flatfield/primary) to get the
> line integral, i.e., the input of most CT reconstruction algorithms.
>
>
> Yep, will do + what I just say. I suggest to continue the discussion when
> something is not clear and I'll summarize soon on the wiki. Keep asking!
>
>
> Thanks,
>
> Lukas
>
> ----------------------------------------------------
> Lukas Gromann
>
> Lehrstuhl für Biomedizinische Physik (E17)
> Technische Universität München
> Boltzmannstrasse 11
> 85748 Garching
>
> Lukas.Gromann at TUM.de
>
> Tel.: 089-289/10905
> Mobil: 017629718306
>
> Am 02.09.2016 um 11:02 schrieb Simon Rit <simon.rit at creatis.insa-lyon.fr>:
>
> Hi,
>
> I assume you've read
>
> http://wiki.opengatecollaboration.org/index.php/Users_Guide_
> V7.2:Readout_parameters_for_Radiotherapy_applications:_Actor
> s#Fixed_Forced_Detection_CT?
>
> It's scarce but there are a bunch of references we suggest to understand
> the technique. The numbers are the expected signal in the image. All
> contributions are summed in total which can be decomposed in
> primary+secondary. Secondary (same as scatter) can itself be decomposed in
> compton+rayleigh+fluorescence. flatfield is available to compute the
> measured primary signal if there is no object, this is useful for CT to
> apply the beer Lambert law. attenuation is ln(flatfield/primary) to get the
> line integral, i.e., the input of most CT reconstruction algorithms.
>
> An alternative is to do pure monte carlo, see monteCarloCT.mac in the same
> folder. I guess there are other options in Gate but I don't know all the CT
> options in Gate. Please let me know if you find some!
>
> Hope this helps and let us know if/how we should complete the "doc",
>
> Simon
>
> On 02/09/2016 09:49, Lukas Gromann wrote:
>
> Dear all,
>
> I am currently learning how to use Gate for my PHD project in medical
> x-rax imaging.
>
> I would like to simulate the Compton scatter fraction in a simple Thorax
> x-ray. I started by using the example_CT/fixedForcedDetectionCT, which
> looks like the thing I am looking for. Unfortunately I cannot find any
> documentation, what is actually the content of the outputfiles created by
> the FixedForcedDetection actor. What does the numbers in the various images
> represent? What is the purpose of the flatfield file?
> Any kind of additional documentation/information would be very welcome!
>
> On the other hand, is there maybe a much better approach to determine
> Compton scatter fractions in my images than using the FFD actor?
>
> Thanks for your help,
>
> Lukas
>
> ----------------------------------------------------
> Lukas Gromann
>
> Lehrstuhl für Biomedizinische Physik (E17)
> Technische Universität München
> Boltzmannstrasse 11
> 85748 Garching
>
> Lukas.Gromann at TUM.de
>
> Tel.: 089-289/10905
> Mobil: 017629718306
>
>
>
> _______________________________________________
> Gate-users mailing listGate-users at lists.opengatecollaboration.orghttp://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
> On 02/09/2016 09:49, Lukas Gromann wrote:
>
> Dear all,
>
> I am currently learning how to use Gate for my PHD project in medical
> x-rax imaging.
>
> I would like to simulate the Compton scatter fraction in a simple Thorax
> x-ray. I started by using the example_CT/fixedForcedDetectionCT, which
> looks like the thing I am looking for. Unfortunately I cannot find any
> documentation, what is actually the content of the outputfiles created by
> the FixedForcedDetection actor. What does the numbers in the various images
> represent? What is the purpose of the flatfield file?
> Any kind of additional documentation/information would be very welcome!
>
> On the other hand, is there maybe a much better approach to determine
> Compton scatter fractions in my images than using the FFD actor?
>
> Thanks for your help,
>
> Lukas
>
> ----------------------------------------------------
> Lukas Gromann
>
> Lehrstuhl für Biomedizinische Physik (E17)
> Technische Universität München
> Boltzmannstrasse 11
> 85748 Garching
>
> Lukas.Gromann at TUM.de
>
> Tel.: 089-289/10905
> Mobil: 017629718306
>
>
>
> _______________________________________________
> Gate-users mailing listGate-users at lists.opengatecollaboration.orghttp://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
> _______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
>
>
> _______________________________________________
> Gate-users mailing listGate-users at lists.opengatecollaboration.orghttp://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
> <responseDetector.txt>_______________________________________________
> Gate-users mailing list
> Gate-users at lists.opengatecollaboration.org
> http://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
>
>
> _______________________________________________
> Gate-users mailing listGate-users at lists.opengatecollaboration.orghttp://lists.opengatecollaboration.org/mailman/listinfo/gate-users
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20160908/14e595f4/attachment-0001.html>