search

rbumm / SlicerLungCTAnalyzer

This is a 3D Slicer extension for segmentation and spatial reconstruction of infiltrated, collapsed, and emphysematous areas in lung CT.
Apache License 2.0
71 stars 23 forks source link

CSV report file #25

Closed romendacil64 closed 3 years ago

romendacil64 commented 3 years ago

Hey there,

I have very recently started using 3D slicer and the Lung Segmentation and Analyzer tool.

I am using nifti-files that for reasons of anonymity do not include metadata about the patient. Unfortunately, in such cases, the csv-report file created by the analyzer is only named with a timestamp - is there a way to change this to timestamp + filename?

Your tool is really fantastic and I am looking forward to work with it!

Best Romendacil

rbumm commented 3 years ago

Thank you @romendacil64 , great that you like the tool. 

Which filename would you need to get included? 

Currently, there are two workarounds  - the easiest one maybe is to open the results folder via the "Open results folder" button directly after the CSV creation, sort the files by time created and add a prefix to the filenames. Secondly, you could call the logic of LungCTAnalyzer by script, this is quite complicated, though time-saving if you need to process a lot of files in a row. For that procedure, you would need to prepare the lung CT datasets including the lung masks, and define the thresholds before running the script.  

Best

Rudolf

lassoan commented 3 years ago

Another option would be to use the input volume node's name as report filename prefix if the volume is not loaded from DICOM.

lassoan commented 3 years ago

@romendacil64 note that nifti is not a general-purpose 3D image file format. It is developed for brain MRI images and should not be used for anything else because it is a very limited file format with many problems (different software interpret nifti files differently, which may cause mismatch in image orientation or may cause the image fail to load, and it has other disadvantages, such as non-human readable/editable/extensible file header). I would recommend to use nrrd file format instead.

rbumm commented 3 years ago

Yes, thanks, I just realized what @romendacil64 wanted. The input volume node name is not ideal because f.e. the public COVID CT dataset uses identical volume names (ct.nrrd) distributed to different foldes ("Pat 1","Pat 2","Pat 3" ...)

rbumm commented 3 years ago

But it is maybe better than just to use a timestamp in anonymized data ?

rbumm commented 3 years ago

Good news, input volume node name is now implemented as a prefix in the CSV report filename.   

@romendacil64 pls wait a few days until the extension manager offers the extension update.

rbumm commented 3 years ago

closed