Registration

classic Classic list List threaded Threaded
26 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Registration

Annie-Pier Lavallée
I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andras Lasso-2

Andriy has implemented a Slicer module for surface based registration and Csaba recently generalized and improved the method. It seems to work very well for prostate contour registration at multiple institutions and probably it would work for other organs. The only problem is that the publicly available registration module in SlicerProstate is very fragile and Csaba’s improvements require SlicerRT and has not been released as a publicly available module yet.

 

Probably the best would be to implement a new, generic (not prostate-specific) “Segment registration” module in SlicerRT, based on Andriy’s original method but including Csaba’s improvements. Probably Csaba could implement it quite quickly, as it would be very similar to his current work-in-progress module.

 

Csaba, Andriy, what do you think?

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Annie-Pier Lavallée
Sent: April 2, 2016 8:17 AM
To: SPL Slicer Users <[hidden email]>
Subject: [slicer-users] Registration

 

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?



thank you

 


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Dominik Meier-2
In reply to this post by Annie-Pier Lavallée
Hi Annie

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:
Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

Best Regards,
Dominik


From: Annie-Pier Lavallée <[hidden email]>
Subject: [slicer-users] Registration
Date: April 2, 2016 at 8:17:17 AM EDT
To: SPL Slicer Users <[hidden email]>
Resent-From: <[hidden email]>

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you




_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Annie-Pier Lavallée
In reply to this post by Annie-Pier Lavallée
Thank a lot for your help! I will try the module surface registration.


Le samedi 2 avril 2016, Dominik Meier <[hidden email]> a écrit :
Hi Annie

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:
Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

Best Regards,
Dominik


From: Annie-Pier Lavallée <<a href="javascript:_e(%7B%7D,'cvml','anniepier.lavallee@gmail.com');" target="_blank">anniepier.lavallee@...>
Subject: [slicer-users] Registration
Date: April 2, 2016 at 8:17:17 AM EDT
To: SPL Slicer Users <<a href="javascript:_e(%7B%7D,'cvml','slicer-users@bwh.harvard.edu');" target="_blank">slicer-users@...>
Resent-From: <<a href="javascript:_e(%7B%7D,'cvml','slicer-users@bwh.harvard.edu');" target="_blank">slicer-users@...>

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you




_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andras Lasso-2
In reply to this post by Dominik Meier-2

As far as I remember that old module used ICP and could only do rigid registration. If rigid transform is acceptable then you can compute that in Slicer4 with a few lines of Python code using this class:

http://www.vtk.org/doc/nightly/html/classvtkIterativeClosestPointTransform.h

 

Binary volumes do not register very well using intensity-based registration methods. Andriy/Csaba’s module compute distance images and registers those using intensity-based transforms. It's more robust and gives meaningful, smooth displacement field in the entire volume, not just at the surface boundaries.

 

Andras

 

From: [hidden email]
Sent: April 2, 2016 16:28
To: [hidden email]
Cc: [hidden email]
Subject: Re: [slicer-users] Registration

 

Hi Annie

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:
Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

Best Regards,
Dominik


From: Annie-Pier Lavallée <[hidden email]>
Subject: [slicer-users] Registration
Date: April 2, 2016 at 8:17:17 AM EDT
To: SPL Slicer Users <[hidden email]>
Resent-From: <[hidden email]>

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you




_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andras Lasso-2
In reply to this post by Annie-Pier Lavallée

Here is a rigid ICP registration module for Slicer4. It requires a small simplification to register two models instead a model and markups.

 

https://github.com/SlicerIGT/SlicerIGT/blob/master/FiducialsToModelRegistration/FiducialsToModelRegistration.py

 

Andras

 

From: [hidden email]
Sent: April 2, 2016 16:36
To: [hidden email]
Cc: [hidden email]
Subject: Re: [slicer-users] Registration

 

Thank a lot for your help! I will try the module surface registration.


Le samedi 2 avril 2016, Dominik Meier <[hidden email]> a écrit :
Hi Annie

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:
Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

Best Regards,
Dominik


From: Annie-Pier Lavallée <anniepier.lavallee@...>
Subject: [slicer-users] Registration
Date: April 2, 2016 at 8:17:17 AM EDT
To: SPL Slicer Users <slicer-users@...>
Resent-From: <slicer-users@...>

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you




_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2
In reply to this post by Annie-Pier Lavallée

I think the suggested plan make a lot of sense.

Annie, what is your input stored in? If it's DICOM-RT, then the surface registration module in the old (>5 years old!) Slicer version 3 would not be very helpful to you, as you cannot import such datasets there.

csaba

On April 2, 2016 15:14:19 Andras Lasso <[hidden email]> wrote:

Andriy has implemented a Slicer module for surface based registration and Csaba recently generalized and improved the method. It seems to work very well for prostate contour registration at multiple institutions and probably it would work for other organs. The only problem is that the publicly available registration module in SlicerProstate is very fragile and Csaba’s improvements require SlicerRT and has not been released as a publicly available module yet.

 

Probably the best would be to implement a new, generic (not prostate-specific) “Segment registration” module in SlicerRT, based on Andriy’s original method but including Csaba’s improvements. Probably Csaba could implement it quite quickly, as it would be very similar to his current work-in-progress module.

 

Csaba, Andriy, what do you think?

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Annie-Pier Lavallée
Sent: April 2, 2016 8:17 AM
To: SPL Slicer Users <[hidden email]>
Subject: [slicer-users] Registration

 

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?



thank you

 


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2
In reply to this post by Annie-Pier Lavallée

Hi Annie,

I think the best would be using the Segment Registration module Andras suggested creating. It would be a quite convenient and robust module to use, and I could add that relatively quickly. We need to come to an agreement first with the other developers though, so please stand by.

csaba

On April 2, 2016 17:43:08 Annie-Pier Lavallée <[hidden email]> wrote:

Yes it's DICOM-RT. So, is it better to load ICP registration module in Slicer 4 with the code that Àndrea send me?

2016-04-02 16:48 GMT-04:00 Csaba Pinter <[hidden email]>:

I think the suggested plan make a lot of sense.

Annie, what is your input stored in? If it's DICOM-RT, then the surface registration module in the old (>5 years old!) Slicer version 3 would not be very helpful to you, as you cannot import such datasets there.

csaba

On April 2, 2016 15:14:19 Andras Lasso <[hidden email]> wrote:

Andriy has implemented a Slicer module for surface based registration and Csaba recently generalized and improved the method. It seems to work very well for prostate contour registration at multiple institutions and probably it would work for other organs. The only problem is that the publicly available registration module in SlicerProstate is very fragile and Csaba’s improvements require SlicerRT and has not been released as a publicly available module yet.

 

Probably the best would be to implement a new, generic (not prostate-specific) “Segment registration” module in SlicerRT, based on Andriy’s original method but including Csaba’s improvements. Probably Csaba could implement it quite quickly, as it would be very similar to his current work-in-progress module.

 

Csaba, Andriy, what do you think?

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Annie-Pier Lavallée
Sent: April 2, 2016 8:17 AM
To: SPL Slicer Users <[hidden email]>
Subject: [slicer-users] Registration

 

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?



thank you

 



_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Tamas Ungi-2
In reply to this post by Andras Lasso-2

I think this one registers two models:

https://github.com/SlicerIGT/SlicerIGT/tree/master/ModelRegistration

 

You get both modules under IGT category, if you install the SlicerIGT extension.

 

Tamas

 

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andras Lasso
Sent: April 2, 2016 4:48 PM
To: Annie-Pier Lavallée <[hidden email]>; Dominik Meier <[hidden email]>
Cc: SPL Slicer Users <[hidden email]>
Subject: Re: [slicer-users] Registration

 

Here is a rigid ICP registration module for Slicer4. It requires a small simplification to register two models instead a model and markups.

 

https://github.com/SlicerIGT/SlicerIGT/blob/master/FiducialsToModelRegistration/FiducialsToModelRegistration.py

 

Andras

 

From: [hidden email]
Sent: April 2, 2016 16:36
To: [hidden email]
Cc: [hidden email]
Subject: Re: [slicer-users] Registration

 

Thank a lot for your help! I will try the module surface registration.

 


Le samedi 2 avril 2016, Dominik Meier <[hidden email]> a écrit :

Hi Annie

 

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:

Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

 

Best Regards,

Dominik

 

 

From: Annie-Pier Lavallée <anniepier.lavallee@...>

Subject: [slicer-users] Registration

Date: April 2, 2016 at 8:17:17 AM EDT

To: SPL Slicer Users <slicer-users@...>

Resent-From: <slicer-users@...>

 

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

 

thank you

 

 

 


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andrey Fedorov-2
In reply to this post by Andras Lasso-2
Annie-Pier - what kind of data are you importing from DICOM? Are these
voxel labels, RT contours, or triangle surfaces?

For the reference, this is the module Andras mentioned, which is
available in the SlicerProstate extension:
http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration

> The only problem is that the publicly
> available registration module in SlicerProstate is very fragile

Andras, can you be more specific what is fragile?

The known issues are that the module expects the same geometry for the
input images [1] and that repeated execution does not work [2]. Are
there any other issues I am not aware of? I was mistakenly mistakenly
thinking Csaba's student was going to fix that problem, and then I
forgot about it. I should fix it.

> Probably the best would be to implement a new, generic (not
> prostate-specific) “Segment registration” module in SlicerRT, based on
> Andriy’s original method but including Csaba’s improvements.

The module we discuss is not prostate specific (and we discussed this
before releasing the module), see documentation [3] -- the module is
called Distance Map Based Registration.

Hard to know what is best to do without knowing what Csaba's
improvements are. In any case, distance map registration module should
and will be fixed to not expect to have the same geometry for the
registered volumes.

[1] https://github.com/SlicerProstate/SlicerProstate/issues/6
[2] https://github.com/SlicerProstate/SlicerProstate/issues/5
[3] http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration

On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:

> Andriy has implemented a Slicer module for surface based registration and
> Csaba recently generalized and improved the method. It seems to work very
> well for prostate contour registration at multiple institutions and probably
> it would work for other organs. The only problem is that the publicly
> available registration module in SlicerProstate is very fragile and Csaba’s
> improvements require SlicerRT and has not been released as a publicly
> available module yet.
>
>
>
> Probably the best would be to implement a new, generic (not
> prostate-specific) “Segment registration” module in SlicerRT, based on
> Andriy’s original method but including Csaba’s improvements. Probably Csaba
> could implement it quite quickly, as it would be very similar to his current
> work-in-progress module.
>
>
>
> Csaba, Andriy, what do you think?
>
>
>
> Andras
>
>
>
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Annie-Pier
> Lavallée
> Sent: April 2, 2016 8:17 AM
> To: SPL Slicer Users <[hidden email]>
> Subject: [slicer-users] Registration
>
>
>
> I have 2 segmentations (2 dicom files) and I want to do a registration with
> the segmentation surface. Is there a module that can do this kind of
> registation?
>
>
>
> thank you
>
>
>
>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email] with
> unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>
>
> The information in this e-mail is intended only for the person to whom it is
> addressed. If you believe this e-mail was sent to you in error and the
> e-mail
> contains patient information, please contact the Partners Compliance
> HelpLine at
> http://www.partners.org/complianceline . If the e-mail was sent to you in
> error
> but does not contain patient information, please contact the sender and
> properly
> dispose of the e-mail.
>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2
HI Andriy,

Annie's data is DICOM-RT.

Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.

My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
1) Support of DICOM, and thus Subject hierarchy, which means transformation of the whole study at once
2) Use of Segmentations infrastructure, thus robust conversion from DICOM-RT to labelmap, which is the input of your module, and on-the fly transformation in any representation with arbitrary transform of the whole structure set (preview of the whole deformed structure set before evaluation/export in both labelmap and surface)
3) Several pre-processing steps (see spacing and area mentioned before, which mean automatic resampling and cropping) that digest the data to yield acceptable results when fed to your module
4) DICOM-RT export of the deformed structure set

These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.

Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.

csaba

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
Sent: April 2, 2016 22:17
To: Andras Lasso
Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
Subject: Re: [slicer-users] Registration

Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?

For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration

> The only problem is that the publicly
> available registration module in SlicerProstate is very fragile

Andras, can you be more specific what is fragile?

The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.

> Probably the best would be to implement a new, generic (not
> prostate-specific) "Segment registration" module in SlicerRT, based on
> Andriy's original method but including Csaba's improvements.

The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.

Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.

[1] https://github.com/SlicerProstate/SlicerProstate/issues/6
[2] https://github.com/SlicerProstate/SlicerProstate/issues/5
[3] http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration

On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:

> Andriy has implemented a Slicer module for surface based registration
> and Csaba recently generalized and improved the method. It seems to
> work very well for prostate contour registration at multiple
> institutions and probably it would work for other organs. The only
> problem is that the publicly available registration module in
> SlicerProstate is very fragile and Csaba's improvements require
> SlicerRT and has not been released as a publicly available module yet.
>
>
>
> Probably the best would be to implement a new, generic (not
> prostate-specific) "Segment registration" module in SlicerRT, based on
> Andriy's original method but including Csaba's improvements. Probably
> Csaba could implement it quite quickly, as it would be very similar to
> his current work-in-progress module.
>
>
>
> Csaba, Andriy, what do you think?
>
>
>
> Andras
>
>
>
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Annie-Pier
> Lavallée
> Sent: April 2, 2016 8:17 AM
> To: SPL Slicer Users <[hidden email]>
> Subject: [slicer-users] Registration
>
>
>
> I have 2 segmentations (2 dicom files) and I want to do a registration
> with the segmentation surface. Is there a module that can do this kind
> of registation?
>
>
>
> thank you
>
>
>
>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>
>
> The information in this e-mail is intended only for the person to whom
> it is addressed. If you believe this e-mail was sent to you in error
> and the e-mail contains patient information, please contact the
> Partners Compliance HelpLine at http://www.partners.org/complianceline 
> . If the e-mail was sent to you in error but does not contain patient
> information, please contact the sender and properly dispose of the
> e-mail.
>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andrey Fedorov-2
In reply to this post by Andrey Fedorov-2
Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working
with DICOM RT (and it seems that she is from Queens U. anyway!), it
definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:

> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from DICOM-RT to labelmap, which is the input of your module, and on-the fly transformation in any representation with arbitrary transform of the whole structure set (preview of the whole deformed structure set before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned before, which mean automatic resampling and cropping) that digest the data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration
>
>> The only problem is that the publicly
>> available registration module in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based on
>> Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3] http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based on
>> Andriy's original method but including Csaba's improvements. Probably
>> Csaba could implement it quite quickly, as it would be very similar to
>> his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>
>> I have 2 segmentations (2 dicom files) and I want to do a registration
>> with the segmentation surface. Is there a module that can do this kind
>> of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>
>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to whom
>> it is addressed. If you believe this e-mail was sent to you in error
>> and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email] with unsubscribe as the subject http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2
In reply to this post by Andrey Fedorov-2
Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:

> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>
>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>
>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 
I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?




2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:
Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>
>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>
>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andrey Fedorov-2
Annie, this coming week I will look into fixing the geometry requirement.

To answer your question, the two modules use completely different algorithms.

I thought I saw your email at one point from Queen’s U. related domain, sorry for the confusion!


On Apr 3, 2016, at 9:06 PM, Csaba Pinter <[hidden email]> wrote:

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 
I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?




2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:
Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>
>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>
>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Andras Lasso-2

Annie, Model registration module in SlicerIGT computes a rigid transform. Most likely you need a deformable transformation. If you can write a bit about your project then we might be able to give specific advice about which method is more appropriate.

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
Sent: April 3, 2016 9:14 PM
To: Csaba Pinter <[hidden email]>
Cc: SPL Slicer Users <[hidden email]>
Subject: Re: [slicer-users] Registration

 

Annie, this coming week I will look into fixing the geometry requirement.

 

To answer your question, the two modules use completely different algorithms.

 

I thought I saw your email at one point from Queen’s U. related domain, sorry for the confusion!

 

 

On Apr 3, 2016, at 9:06 PM, Csaba Pinter <[hidden email]> wrote:

 

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 

I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?

 

 

 

2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:

Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>

>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>

>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ

 

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

 


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Annie-Pier Lavallée
In reply to this post by Andrey Fedorov-2
Ok, thank ofr your help. For my project, I need a rigid transfomation so Model Registration seems to be the best module.

2016-04-03 22:30 GMT-04:00 Andras Lasso <[hidden email]>:

Annie, Model registration module in SlicerIGT computes a rigid transform. Most likely you need a deformable transformation. If you can write a bit about your project then we might be able to give specific advice about which method is more appropriate.

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
Sent: April 3, 2016 9:14 PM
To: Csaba Pinter <[hidden email]>
Cc: SPL Slicer Users <[hidden email]>
Subject: Re: [slicer-users] Registration

 

Annie, this coming week I will look into fixing the geometry requirement.

 

To answer your question, the two modules use completely different algorithms.

 

I thought I saw your email at one point from Queen’s U. related domain, sorry for the confusion!

 

 

On Apr 3, 2016, at 9:06 PM, Csaba Pinter <[hidden email]> wrote:

 

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 

I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?

 

 

 

2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:

Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>

>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>

>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ

 

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

 



_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Dominik Meier-2
In reply to this post by Annie-Pier Lavallée
Hi Andras

Many thanks!  Very useful as this is a fairly common request. I added the link to the FAQ.
Best,
Dominik


On 02.04.2016, at 22:48, Andras Lasso <[hidden email]> wrote:

Here is a rigid ICP registration module for Slicer4. It requires a small simplification to register two models instead a model and markups.

 

 

Andras

 

From: [hidden email]
Sent: April 2, 2016 16:36
To: [hidden email]
Cc: [hidden email]
Subject: Re: [slicer-users] Registration

 

Thank a lot for your help! I will try the module surface registration.


Le samedi 2 avril 2016, Dominik Meier <[hidden email]> a écrit :
Hi Annie

There used to be a surface registration in earlier versions, which has not (yet) been updated to the latest release. Details and links here:
Otherwise if your segmentations cover a sufficient volume, you may be able to run a regular registration with the segmentation volumes directly, or preferably with the original images using the segmentations as a mask.

Best Regards,
Dominik


From: Annie-Pier Lavallée <anniepier.lavallee@...>
Subject: [slicer-users] Registration
Date: April 2, 2016 at 8:17:17 AM EDT
To: SPL Slicer Users <slicer-users@...>
Resent-From: <slicer-users@...>

I have 2 segmentations (2 dicom files) and I want to do a registration with the segmentation surface. Is there a module that can do this kind of registation?

thank you


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Csaba Pinter-2
In reply to this post by Annie-Pier Lavallée

Hi Annie,

 

I quickly created a simplified registration module that you can try. You can of course use ModelRegistration, but for that you’ll need to do some data juggling. Let us know if you get stuck with that.

 

The new module is called SegmentRegistration, and you can download it from here [1]. You’ll need a recent nightly [2], on top of which you need to install SlicerRT and SlicerProstate from the Extension Manager. After that, add the folder where you put SegmentRegistration.py in the Additional module paths section on the Modules pane of Application Settings.

 

csaba

 

[1] https://subversion.assembla.com/svn/slicerrt/trunk/SlicerRt/sandbox/ProstateMRIUSContourPropagation/SegmentRegistration.py

[2] http://slicer.cdash.org/index.php?project=Slicer4 (click the yellow package button for your platform)

  Quick link for windows: http://slicer.kitware.com/midas3/api/rest?method=midas.bitstream.download&name=Slicer-4.5.0-2016-04-01-win-amd64.exe&checksum=05936067d26ea442be9834311fef6a72

 

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Annie-Pier Lavallée
Sent: April 3, 2016 22:39
To: Andras Lasso
Cc: SPL Slicer Users
Subject: Re: [slicer-users] Registration

 

Ok, thank ofr your help. For my project, I need a rigid transfomation so Model Registration seems to be the best module.

 

2016-04-03 22:30 GMT-04:00 Andras Lasso <[hidden email]>:

Annie, Model registration module in SlicerIGT computes a rigid transform. Most likely you need a deformable transformation. If you can write a bit about your project then we might be able to give specific advice about which method is more appropriate.

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
Sent: April 3, 2016 9:14 PM
To: Csaba Pinter <[hidden email]>
Cc: SPL Slicer Users <[hidden email]>
Subject: Re: [slicer-users] Registration

 

Annie, this coming week I will look into fixing the geometry requirement.

 

To answer your question, the two modules use completely different algorithms.

 

I thought I saw your email at one point from Queen’s U. related domain, sorry for the confusion!

 

 

On Apr 3, 2016, at 9:06 PM, Csaba Pinter <[hidden email]> wrote:

 

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 

I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?

 

 

 

2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:

Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>

>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>

>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ

 

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

 

 


_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
Reply | Threaded
Open this post in threaded view
|

Re: Registration

Annie-Pier Lavallée
In reply to this post by Annie-Pier Lavallée
Thank a lot. I tried your new module. The registration was not as good as with the ModelRegistration module (see picture).

2016-04-04 11:04 GMT-04:00 Csaba Pinter <[hidden email]>:

Hi Annie,

 

I quickly created a simplified registration module that you can try. You can of course use ModelRegistration, but for that you’ll need to do some data juggling. Let us know if you get stuck with that.

 

The new module is called SegmentRegistration, and you can download it from here [1]. You’ll need a recent nightly [2], on top of which you need to install SlicerRT and SlicerProstate from the Extension Manager. After that, add the folder where you put SegmentRegistration.py in the Additional module paths section on the Modules pane of Application Settings.

 

csaba

 

[1] https://subversion.assembla.com/svn/slicerrt/trunk/SlicerRt/sandbox/ProstateMRIUSContourPropagation/SegmentRegistration.py

[2] http://slicer.cdash.org/index.php?project=Slicer4 (click the yellow package button for your platform)

  Quick link for windows: http://slicer.kitware.com/midas3/api/rest?method=midas.bitstream.download&name=Slicer-4.5.0-2016-04-01-win-amd64.exe&checksum=05936067d26ea442be9834311fef6a72

 

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Annie-Pier Lavallée
Sent: April 3, 2016 22:39
To: Andras Lasso
Cc: SPL Slicer Users
Subject: Re: [slicer-users] Registration

 

Ok, thank ofr your help. For my project, I need a rigid transfomation so Model Registration seems to be the best module.

 

2016-04-03 22:30 GMT-04:00 Andras Lasso <[hidden email]>:

Annie, Model registration module in SlicerIGT computes a rigid transform. Most likely you need a deformable transformation. If you can write a bit about your project then we might be able to give specific advice about which method is more appropriate.

 

Andras

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andrey Fedorov
Sent: April 3, 2016 9:14 PM
To: Csaba Pinter <[hidden email]>
Cc: SPL Slicer Users <[hidden email]>
Subject: Re: [slicer-users] Registration

 

Annie, this coming week I will look into fixing the geometry requirement.

 

To answer your question, the two modules use completely different algorithms.

 

I thought I saw your email at one point from Queen’s U. related domain, sorry for the confusion!

 

 

On Apr 3, 2016, at 9:06 PM, Csaba Pinter <[hidden email]> wrote:

 

Hi Annie,

I cannot help you with either module, so let's keep the mailing list in the loop.

I'm trying to arrange the best solution for your problem, but if you need results right now, then you'll need to ask the module owners (Andriy and Tamas respectively) for help.

csaba

On April 3, 2016 8:47:48 PM Annie-Pier Lavallée <[hidden email]> wrote:

I'm at Polytechnique Montreal. 

I transformed my DICOM RT in labelmap volume and I tried to use DistanceMapBasedRegistration but it did'nt work. The transforms matrix were empty. Instead, I used Model registration in IGT extension. Is both modules use a similar algorithm?

 

 

 

2016-04-03 19:34 GMT-04:00 Csaba Pinter <[hidden email]>:

Hi Andriy,

We have a very application-specific module that does the registration on segmentations. If she or someone else wants to do this workflow, then it would be needed to generalize it, which would not be a big work. But there are a few questions about how to do that. Such as do we still directly call DistanceMapBasedRegistration from the SlicerProstate extension to run a not prostate specific algorithm? Or do we keep using that code and do pre- and post-processing, or integrate parts of it in the new module?

Annie, are you really at Queen's University as Andriy says?

Cheers,
csaba

-----Original Message-----
From: Andrey Fedorov [mailto:[hidden email]]
Sent: April 3, 2016 14:50
To: Csaba Pinter
Cc: Andras Lasso; SPL Slicer Users
Subject: Re: [slicer-users] Registration

Csaba,

thanks for the details.

Since all of this has already been done, and Annie-Pier is working with DICOM RT (and it seems that she is from Queens U. anyway!), it definitely makes sense for her to use your module.

AF

On Sun, Apr 3, 2016 at 11:17 AM, Csaba Pinter <[hidden email]> wrote:
> HI Andriy,
>
> Annie's data is DICOM-RT.
>
> Fragility means that it works with many assumptions. Besides the two issues you mention I reported, the most important is that it also assumes the datasets to have voxel spacings in a specific range, and that the structures need to be in a certain area of the input labelmaps. Also there are other minor issues, such as one of the result models does not appear, and an issue with CLI execution.
>
> My improvements in the workflow-based (but not slicelet) module I created for a clinical project:
> 1) Support of DICOM, and thus Subject hierarchy, which means
> transformation of the whole study at once
> 2) Use of Segmentations infrastructure, thus robust conversion from
> DICOM-RT to labelmap, which is the input of your module, and on-the
> fly transformation in any representation with arbitrary transform of
> the whole structure set (preview of the whole deformed structure set
> before evaluation/export in both labelmap and surface)
> 3) Several pre-processing steps (see spacing and area mentioned
> before, which mean automatic resampling and cropping) that digest the
> data to yield acceptable results when fed to your module
> 4) DICOM-RT export of the deformed structure set
>
> These improvements do not mean that it will be surely robust with every dataset, but it does mean that it works now with both the sample datasets you provided and with the clinical data we use for our project, and that it's quite easy-to-use from the user point of view.
>
> Although I cannot be fully objective with the Segmentations infrastructure, I think its advantages are huge and cannot be ignored. Besides the already mentioned bulk management (conversion, visualization, transforms, storage), there are many reliable and stress-tested convenience functions for resampling and other geometrical operations. So these advantages don't just mean robust operation and nice features, but also incomparably reduced development time and number of lines of application code.
> This is why I support Andras' idea to create a segmentations-enabled module for this task. Your distance map based registration module works nicely in general aside from those bugs, this is why we use it even if we need to do various preprocessing steps, but as segmentations is coming to Slicer with its full power really soon, I see this module as a natural step forward.
>
> csaba
>
> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Andrey
> Fedorov
> Sent: April 2, 2016 22:17
> To: Andras Lasso
> Cc: SPL Slicer Users; Fedorov, Andriy Y., Ph.D.
> Subject: Re: [slicer-users] Registration
>
> Annie-Pier - what kind of data are you importing from DICOM? Are these voxel labels, RT contours, or triangle surfaces?
>
> For the reference, this is the module Andras mentioned, which is available in the SlicerProstate extension:
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
>> The only problem is that the publicly available registration module
>> in SlicerProstate is very fragile
>
> Andras, can you be more specific what is fragile?
>
> The known issues are that the module expects the same geometry for the input images [1] and that repeated execution does not work [2]. Are there any other issues I am not aware of? I was mistakenly mistakenly thinking Csaba's student was going to fix that problem, and then I forgot about it. I should fix it.
>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>
> The module we discuss is not prostate specific (and we discussed this before releasing the module), see documentation [3] -- the module is called Distance Map Based Registration.
>
> Hard to know what is best to do without knowing what Csaba's improvements are. In any case, distance map registration module should and will be fixed to not expect to have the same geometry for the registered volumes.
>
> [1] https://github.com/SlicerProstate/SlicerProstate/issues/6
> [2] https://github.com/SlicerProstate/SlicerProstate/issues/5
> [3]
> http://wiki.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modu
> les/DistanceMapBasedRegistration
>
> On Sat, Apr 2, 2016 at 3:14 PM, Andras Lasso <[hidden email]> wrote:
>> Andriy has implemented a Slicer module for surface based registration
>> and Csaba recently generalized and improved the method. It seems to
>> work very well for prostate contour registration at multiple
>> institutions and probably it would work for other organs. The only
>> problem is that the publicly available registration module in
>> SlicerProstate is very fragile and Csaba's improvements require
>> SlicerRT and has not been released as a publicly available module yet.
>>
>>
>>
>> Probably the best would be to implement a new, generic (not
>> prostate-specific) "Segment registration" module in SlicerRT, based
>> on Andriy's original method but including Csaba's improvements.
>> Probably Csaba could implement it quite quickly, as it would be very
>> similar to his current work-in-progress module.
>>
>>
>>
>> Csaba, Andriy, what do you think?
>>
>>
>>
>> Andras
>>
>>
>>
>> From: [hidden email]
>> [mailto:[hidden email]] On Behalf Of Annie-Pier
>> Lavallée
>> Sent: April 2, 2016 8:17 AM
>> To: SPL Slicer Users <[hidden email]>
>> Subject: [slicer-users] Registration
>>
>>
>>

>> I have 2 segmentations (2 dicom files) and I want to do a
>> registration with the segmentation surface. Is there a module that
>> can do this kind of registation?
>>
>>
>>
>> thank you
>>
>>
>>
>>

>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to [hidden email]
>> with unsubscribe as the subject
>> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
>>
>>
>> The information in this e-mail is intended only for the person to
>> whom it is addressed. If you believe this e-mail was sent to you in
>> error and the e-mail contains patient information, please contact the
>> Partners Compliance HelpLine at
>> http://www.partners.org/complianceline
>> . If the e-mail was sent to you in error but does not contain patient
>> information, please contact the sender and properly dispose of the
>> e-mail.
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email]
> with unsubscribe as the subject
> http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ

 

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

 

 



_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
http://www.slicer.org/slicerWiki/index.php/Documentation/4.3/FAQ


The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.

NewRegistration.JPG (179K) Download Attachment
12