Large data segmentation

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

Large data segmentation

romiplitman
Hi All,

Due to bad resolution in the sagittal and coronal planes of my confocal microscopic images, I had to resample the data. This leaded to a very large data (around 200 slices). Is there a way to do a quick segmentation? My problem is that my images are complicated (blurred) so automatic segmentation is problematic. I usually apply a threshold effect and then arrange each slice manually by cleaning the contours and adding missed areas. I know that in some commercial softwares there is the possibility to segment 2 slices and then do interpolation to those slices in between, this is very useful in my case, is there something similar in Slicer? Any ideas?

Thanks,

Romina
Reply | Threaded
Open this post in threaded view
|

Re: Large data segmentation

Steve Pieper-2
Hi Romina - 

You might try the CarreraSlice extension in the nightly build [1].  

-Steve

[1] https://www.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/CarreraSliceInteractiveSegmenter


On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
Hi All,

Due to bad resolution in the sagittal and coronal planes of my confocal
microscopic images, I had to resample the data. This leaded to a very large
data (around 200 slices). Is there a way to do a quick segmentation? My
problem is that my images are complicated (blurred) so automatic
segmentation is problematic. I usually apply a threshold effect and then
arrange each slice manually by cleaning the contours and adding missed
areas. I know that in some commercial softwares there is the possibility to
segment 2 slices and then do interpolation to those slices in between, this
is very useful in my case, is there something similar in Slicer? Any ideas?

Thanks,

Romina



--
View this message in context: http://slicer-users.65878.n3.nabble.com/Large-data-segmentation-tp4027293.html
Sent from the slicer-users mailing list archive at Nabble.com.
_______________________________________________
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: Large data segmentation

Robert Funnell
Romina -

If I understand correctly, you had fewer slices to start with but
ended up with 200 after supersampling the volume. If that's so, why
not just supersample the individual slices?

- Robert

> On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
>
>> Hi All,
>>
>> Due to bad resolution in the sagittal and coronal planes of my confocal
>> microscopic images, I had to resample the data. This leaded to a very large
>> data (around 200 slices). Is there a way to do a quick segmentation? My
>> problem is that my images are complicated (blurred) so automatic
>> segmentation is problematic. I usually apply a threshold effect and then
>> arrange each slice manually by cleaning the contours and adding missed
>> areas. I know that in some commercial softwares there is the possibility to
>> segment 2 slices and then do interpolation to those slices in between, this
>> is very useful in my case, is there something similar in Slicer? Any ideas?
>>
>> Thanks,
>>
>> Romina
>>
_______________________________________________
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: Large data segmentation

Robert Funnell
Romina -

I don't know how to do it in Slicer, but I would get the individual
slice images into individual image files and then use GIMP or
ImageMagick to resample them, and then get them back into Slicer. So
the number of slices remains the same.

- Robert

On Tue, 4 Mar 2014, Romina Plitman wrote:

> Hi Robert,
>
> I am not sure I understood what you mean by supersample the individual
> slices? ...
>
>
> On Tue, Mar 4, 2014 at 12:44 PM, Robert Funnell <[hidden email]>wrote:
>
>> Romina -
>>
>> If I understand correctly, you had fewer slices to start with but ended up
>> with 200 after supersampling the volume. If that's so, why not just
>> supersample the individual slices?
>>
>> - Robert
>>
>>
>>  On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
>>>
>>>  Hi All,
>>>>
>>>> Due to bad resolution in the sagittal and coronal planes of my confocal
>>>> microscopic images, I had to resample the data. This leaded to a very
>>>> large
>>>> data (around 200 slices). Is there a way to do a quick segmentation? My
>>>> problem is that my images are complicated (blurred) so automatic
>>>> segmentation is problematic. I usually apply a threshold effect and then
>>>> arrange each slice manually by cleaning the contours and adding missed
>>>> areas. I know that in some commercial softwares there is the possibility
>>>> to
>>>> segment 2 slices and then do interpolation to those slices in between,
>>>> this
>>>> is very useful in my case, is there something similar in Slicer? Any
>>>> ideas?
>>>>
>>>> Thanks,
>>>>
>>>> Romina
>>>>
_______________________________________________
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: Large data segmentation

Dominik Meier
In reply to this post by Steve Pieper-2
Hi Romina

Carrera Slice is nice. ;-) I had some good results with it. Definitely worth a try.
Upsampling depends very much on how severe your anisotropy is, i.e. how large your slice thickness is relative to the in-plane voxel size, and how important the shape of the 3D segmented structure in that direction. So I would experiment with different upsampling ratios.You might also consider segmenting on a medium size set first and then upsample that to get a smoother outline, there's the "Labelmap Smoothing" module and several filter modules that you can use to reduce the "wedding cake" effect you get from segmentations of very thick slices. You'd first convert the segmentation back into a regular volume (turn off the labelmap checkbox in the Volumes module), then apply a filter (e.g. Gaussian), and then threshold the filtered volume again to get a smoother interpolation. The Labelmap Smoothing module does some of that for you, but if you need more control you may have to break it down into indiv. steps.  This works best if the curvature orthogonal to the slice direction is small.   If you have to segment manually anyway then upsampling is not going to help you much, since the underlying data remains the same. 

Best,
Dominik


On Mar 4, 2014, at 7:34 AM, Steve Pieper <[hidden email]> wrote:

Hi Romina - 

You might try the CarreraSlice extension in the nightly build [1].  

-Steve

[1] https://www.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/CarreraSliceInteractiveSegmenter


On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
Hi All,

Due to bad resolution in the sagittal and coronal planes of my confocal
microscopic images, I had to resample the data. This leaded to a very large
data (around 200 slices). Is there a way to do a quick segmentation? My
problem is that my images are complicated (blurred) so automatic
segmentation is problematic. I usually apply a threshold effect and then
arrange each slice manually by cleaning the contours and adding missed
areas. I know that in some commercial softwares there is the possibility to
segment 2 slices and then do interpolation to those slices in between, this
is very useful in my case, is there something similar in Slicer? Any ideas?

Thanks,

Romina



--
View this message in context: http://slicer-users.65878.n3.nabble.com/Large-data-segmentation-tp4027293.html
Sent from the slicer-users mailing list archive at Nabble.com.
_______________________________________________
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


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.

signature.asc (507 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Large data segmentation

romiplitman
Hi Dominik,

Thank you for your answer. I am trying to install Carrera Slice, however I am unable to see the Extensions Manager. When I open it, its empty ... This happens only in the nightly version (I work in a 64bit linux, so it should be fine). I have also the stable version installed and the extensions manager works fine there, but Carrera Slice module is not there ! :( . Do you know why is this?

Another short question, can I resample a label? If yes, maybe that is a better idea ....

Thx,

Romina
Reply | Threaded
Open this post in threaded view
|

Re: Large data segmentation

Dominik Meier
In reply to this post by Dominik Meier
Yes I think Carrera came after the stable release, and there may have been Extension Manager issues in the very recent build. Some of the developers should be able to help with that.
Yes you can resample a label. There's several resampling modules (Resample Volume, Resample Scalar/Vector/DWI, BRAINs resample).  The important thing to remember there is to select "Nearest Neighbor Interpolation". You can specify the resolution you want directly or indirectly by providing a reference volume from which to copy the desired resolution. The latter tends to be more robust. 

-- Dominik




On Mar 4, 2014, at 9:32 AM, Romina Plitman <[hidden email]> wrote:

Hi Dominik,

Thank you for your answer. I am trying to install Carrera Slice, however I am unable to see the Extensions Manager. When I open it, its empty ... This happens only in the nightly version (I work in a 64bit linux, so it should be fine). I have also the stable version installed and the extensions manager works fine there, but Carrera Slice module is not there ! :( . Do you know why is this?

Another short question, can I resample a label? If yes, maybe that is a better idea ....

Thx,

Romina


On Tue, Mar 4, 2014 at 2:19 PM, Dominik Meier <[hidden email]> wrote:
Hi Romina

Carrera Slice is nice. ;-) I had some good results with it. Definitely worth a try.
Upsampling depends very much on how severe your anisotropy is, i.e. how large your slice thickness is relative to the in-plane voxel size, and how important the shape of the 3D segmented structure in that direction. So I would experiment with different upsampling ratios.You might also consider segmenting on a medium size set first and then upsample that to get a smoother outline, there's the "Labelmap Smoothing" module and several filter modules that you can use to reduce the "wedding cake" effect you get from segmentations of very thick slices. You'd first convert the segmentation back into a regular volume (turn off the labelmap checkbox in the Volumes module), then apply a filter (e.g. Gaussian), and then threshold the filtered volume again to get a smoother interpolation. The Labelmap Smoothing module does some of that for you, but if you need more control you may have to break it down into indiv. steps.  This works best if the curvature orthogonal to the slice direction is small.   If you have to segment manually anyway then upsampling is not going to help you much, since the underlying data remains the same. 

Best,
Dominik


On Mar 4, 2014, at 7:34 AM, Steve Pieper <[hidden email]> wrote:

Hi Romina - 

You might try the CarreraSlice extension in the nightly build [1].  

-Steve

[1] https://www.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/CarreraSliceInteractiveSegmenter


On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
Hi All,

Due to bad resolution in the sagittal and coronal planes of my confocal
microscopic images, I had to resample the data. This leaded to a very large
data (around 200 slices). Is there a way to do a quick segmentation? My
problem is that my images are complicated (blurred) so automatic
segmentation is problematic. I usually apply a threshold effect and then
arrange each slice manually by cleaning the contours and adding missed
areas. I know that in some commercial softwares there is the possibility to
segment 2 slices and then do interpolation to those slices in between, this
is very useful in my case, is there something similar in Slicer? Any ideas?

Thanks,

Romina



--
View this message in context: http://slicer-users.65878.n3.nabble.com/Large-data-segmentation-tp4027293.html
Sent from the slicer-users mailing list archive at Nabble.com.
_______________________________________________
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




--
Romina Plitman Mayo
PhD Candidate
Centre for Trophoblast Research
Department of Physiology, Development and Neuroscience
University of Cambridge
Cambridge, UK


_______________________________________________
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.

signature.asc (507 bytes) Download Attachment
Viz
Reply | Threaded
Open this post in threaded view
|

Re: Large data segmentation

Viz
In reply to this post by romiplitman
Hi Romina,

Do you have a Mac computer around? I know that Carrera works on Mac. Even I had trouble obtaining Carrera on Windows.

Best,
Viz.
Reply | Threaded
Open this post in threaded view
|

Fwd: Large data segmentation

Steve Pieper-2
In reply to this post by Steve Pieper-2
adding the list for the record.

---------- Forwarded message ----------
From: Ivan Kolesov <[hidden email]>
Date: Fri, Mar 7, 2014 at 12:20 PM
Subject: Re: [slicer-users] Large data segmentation
To: Steve Pieper <[hidden email]>
Cc: Romina Plitman <[hidden email]>


Hi Romina,
    CarreraSlice contains all functionality of KSlice in addition to a fast implementation of GrowCut for initialization. Further, there were a number of bug fixes applied as well as tweaks to the user interface to arrive at CarreraSlice. Thus, KSlice is deprecated and is no longer part of Slicer extensions. CarreraSlice also gives you an option to segment slice by slice or in 3D; it sounds like this is something you feel is necessary for your dataset. Perhaps this way you can avoid resampling to 200 slices and segment the original data first, then try to smooth the results, if needed. Let me know if you have more questions.

Best,
   Ivan


On Tue, Mar 4, 2014 at 7:49 AM, Steve Pieper <[hidden email]> wrote:
Hi Ivan - 

Can you give some ideas to Romina about CarreraSlice vs KSlice for her data?

-Steve


On Tue, Mar 4, 2014 at 7:38 AM, Romina Plitman <[hidden email]> wrote:
Hi Steve,

Thank you for your answer. I have tried the Kslicer and although it helps, the segmentation is still slow. Are the Kslicer and Carrera much different?

Thx,

Romina


On Tue, Mar 4, 2014 at 12:34 PM, Steve Pieper <[hidden email]> wrote:
Hi Romina - 

You might try the CarreraSlice extension in the nightly build [1].  

-Steve

[1] https://www.slicer.org/slicerWiki/index.php/Documentation/Nightly/Modules/CarreraSliceInteractiveSegmenter


On Tue, Mar 4, 2014 at 6:10 AM, romiplitman <[hidden email]> wrote:
Hi All,

Due to bad resolution in the sagittal and coronal planes of my confocal
microscopic images, I had to resample the data. This leaded to a very large
data (around 200 slices). Is there a way to do a quick segmentation? My
problem is that my images are complicated (blurred) so automatic
segmentation is problematic. I usually apply a threshold effect and then
arrange each slice manually by cleaning the contours and adding missed
areas. I know that in some commercial softwares there is the possibility to
segment 2 slices and then do interpolation to those slices in between, this
is very useful in my case, is there something similar in Slicer? Any ideas?

Thanks,

Romina



--
View this message in context: http://slicer-users.65878.n3.nabble.com/Large-data-segmentation-tp4027293.html
Sent from the slicer-users mailing list archive at Nabble.com.
_______________________________________________
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.





--
Romina Plitman Mayo
PhD Candidate
Centre for Trophoblast Research
Department of Physiology, Development and Neuroscience
University of Cambridge
Cambridge, UK




_______________________________________________
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.