tbaudier [Thu, 20 Dec 2018 12:08:55 +0000 (13:08 +0100)]
Change dicom file sorting procedure in clitkDicom2Image
In clitkDicom2Image, the sorting of files was based on the slice position but it's not correct when a transformation matrix is set.
It's ok when you open a dicom serie with vv. So I used GDCMSeriesFileNames like with vv to sort filenames correctly
tbaudier [Thu, 20 Dec 2018 10:01:48 +0000 (11:01 +0100)]
Change serie number to serie UID
In clitkDicom2Image, change serie number to serie UID.
Serie UID is more restrictive and better to separate series
Change number to string because series UID is a string
Series UID is better to use itk::GDCMSeriesFileNames
tbaudier [Wed, 19 Dec 2018 14:50:37 +0000 (15:50 +0100)]
Debug opening dicom with rotation matrix
With the introduction of the choose in the Image reader to open the dicom with or without the patient coordinate system, I need to pass the value of the patient flag to take into account the transformation matrix
tbaudier [Thu, 22 Nov 2018 15:36:11 +0000 (16:36 +0100)]
Debug: correct wrong flag in condition
When the flag patientSystem was introduced in clitkDicom2Image to save the patient matrix, the flag to check the tool has to compute this matrix was not correct
tbaudier [Thu, 15 Nov 2018 15:50:13 +0000 (16:50 +0100)]
Check if mask and input have the same spacing to compute SUVPeak
Before, the output of the SUVPeak was unrealistic with a mask with a spacing different from the input.
Now, check if they have the same spacing. If not, the script raises an error.
I add an option (--allow_resize) to automatically resize the mask according to the input like in the clitkImageStatistics script
tbaudier [Thu, 15 Nov 2018 15:24:18 +0000 (16:24 +0100)]
Change behavior of clitkCropImage with like option
Before, if we cropped an image like another, the output's origin was set at the origin of the input.
But we prefer to keep the registration correctly, so we prefer to have an output's origin defined to still have the registration between the output and the input.
This is the normal behavior of the itk::PasteImageFilter.
I removed the SetNumberOfRequiredInputs to avoid this error: PasteImageFilter(0x1927770): At least 2 of the first 2 indexed inputs are required but only 1 are specified. The required inputs are expected to be the first inputs.
I add the updateOrigin flag to keep the previous behavior. If the flag is set to On (default is off), the origin of the output is set to the same origin than the input.
I update the description of the filte about this origin managment
tbaudier [Wed, 14 Nov 2018 12:28:25 +0000 (13:28 +0100)]
Merge dose by region uncertainties using Gate uncertainty calculation
Instead of using partial uncertainty, I use the total dose or edep and the square of the dose and edep like
I removed a bash function because it's useless
tbaudier [Fri, 9 Nov 2018 08:49:28 +0000 (09:49 +0100)]
Modify mergeDoseByRegion
Before, it summed 2 files to write the result.
But I had to modify it to be able to sum partial std. So I created 3 functions:
-addToPartialResult with the previous code
- addWithoutPartialResult for the first file
- divideUncertaintyResult for the last step
tbaudier [Tue, 16 Oct 2018 08:41:26 +0000 (10:41 +0200)]
Add like option in clitkExtrude tool
Before we had to set size, spacing and origin for the new dimension
But we could want to have the same size, spacing and origin than another image.
So I add the like option
I remove the required for the size option because if like is set , size is not mandatory
I controled that like and origin, size or spacing are not set at the same time
Add AffineRegistration library to avoid compilation error
With a lot of cores, clitkElastixTransformToMatrix (that need clitkAffineTransform) can be compiled before clitkAffineTransform
And the gengetopt of clitkAffineTransform is not compiled leading to an error