Thank you very much Stefan,
I'm comftable enough to run a python script indeed and it worked very well.
Thank you so much!
Bernard
Le jeudi 13 février 2025 à 15:34:56 UTC+1, Dinkelacker, Stefan
<[email protected]> a écrit :
Thank you for the image! Do you feel comfortable executing Python scripts? In
that case I asked ChatGPT to create a script to convert DICOM images like yours
to a single channel grayscale image which you then can easily use in MITK with
the Statistics plugin. I attach the script. As prerequisite, you need to
install SimpleITK and numpy with pip: "pip install SimpleITK numpy". Then,
execute the script like:
python convert_dicom_grayscale.py P1FA4OOE
which will create an image file at the same location named
P1FA4OOE_grayscale.nrrd.
Best,
Stefan
________________________________________
Von: Delvaux Bernard <[email protected]>
Gesendet: Mittwoch, 12. Februar 2025 12:59
An: [email protected]; Dinkelacker, Stefan
Betreff: Re: AW: AW: [Extern] - [mitk-users] Retrieve segmented surfaces
Thank you Stefan for your kind answer.
As far as I know it is a raw DICOM file extracted from US machine Logic E10s
(GE Healthcare).
The file is totally anonymized and I will send you a link to download it.
Regards,
Bernard
Le mardi 11 février 2025 à 23:41:20 UTC+1, Dinkelacker, Stefan
<[email protected]> a écrit :
I see, I will extend the ticket to not only handle unsupported pixel types more
gracefully but to maybe also still show statistics that do not depend on the
pixel type like volume and number of pixels of labels. I am afraid none of this
will resolve your issue in the short term since code changes are required. The
empty pixel type is interesting, can you tell us more about the image or in
case it is anonymized give us a sample so we can look up what is going on?
Best,
Stefan
________________________________________
Von: Delvaux Bernard <[email protected]<mailto:[email protected]>>
Gesendet: Dienstag, 11. Februar 2025 19:43
An: [email protected]; Dinkelacker, Stefan
Betreff: Re: AW: [Extern] - [mitk-users] Retrieve segmented surfaces
Hi again Stefan, and thank you again for your help.
I want to measure the surface of a label indeed. And more precisely, calculate
metrics (like DSC, IoU, precision and recall) of 2 labels in a single slice
(for now; we consider expanding to 3d meassurements later).
I already used the method you recommended with Horos: I measured by hand the 2
surfaces and their overlap by hand and I was able to calculate everything I
wanted. So it is possible but time consuming, and also less precise than
automated calculations.
This is why I'd like to use a software with segmentation capabilities for that
purpose.
Thanks
Bernard
Le mardi 11 février 2025 à 10:56:54 UTC+1, Dinkelacker, Stefan
<[email protected]<mailto:[email protected]>> a
écrit :
Sorry, sent again to include the MITK Users list.
________________________________________
Von: Dinkelacker, Stefan
Gesendet: Dienstag, 11. Februar 2025 10:54:43
An: Delvaux Bernard
Betreff: AW: [Extern] - [mitk-users] Retrieve segmented surfaces
Hi again,
by "surface", do you mean a segmentation label, resp. masked pixels and only in
a single slice? Or all the pixels of a segmentation label (volume)? Or, surface
in the sense of MITK, which would be a 3-d polygon mesh?
The volume of segmentation labels is shown in the statistics plugin, but I
already read that you have trouble with unsupported pixel types, right? If you
are really just interested in the 2-d area of something in an image, and you
just created segmentations for this purpose, an alternative would be the
Measurement plugin, where you can draw a (smoothed) polygon for example to see
the circumference and area in mm².
Best,
Stefan
________________________________________
Von: Delvaux Bernard via mitk-users
<[email protected]<mailto:[email protected]><mailto:[email protected]>>
Gesendet: Montag, 10. Februar 2025 07:54
An:
[email protected]<mailto:[email protected]><mailto:[email protected]>
Betreff: [Extern] - [mitk-users] Retrieve segmented surfaces
Hello,
I'd like to know please how to retrieve the area of segmented surfaces,
including intersection and union surfaces calculated in Boolean operations.
I've been searching for this without succes and I'm new ti MITK Workbench.
Thanks,
Bernard
_______________________________________________
mitk-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mitk-users
