jpr [Mon, 23 Oct 2006 15:45:19 +0000 (15:45 +0000)]
Some ACR readers *need* the DataElement (0x0008,0x0000), and cannot recognize
gdcm-written ACR files :-(
DataElement (0x0008,0x0000) is now added to ACR-files.
(I guess nobody cares, but people from Cardiological Hospital,
using *very* old clinical Softwares)
jpr [Wed, 18 Oct 2006 13:40:18 +0000 (13:40 +0000)]
FileHelper::CheckMandatoryElements() now checks as well
Pixel Representation (0028_0103) and defailts it as 0 (unsigned) if missing
to avoid unexpected behavior for images written by unaware users.
" Extracts user supplied tags (identified by group number-element number)",
" and writes their values in a text file. ",
" (file will be used as input by BrainVisa software ) ",
" - Converts the Siemens Sonata MRI '*tfl2d1' DICOM files ",
" to be processable by MriRegister software ",
" - May be used as a template for gdcm::SerieHelper use. ",
Sometimes, two values are *not separated by a white space.
Fortunately, the second one is negative.
I had better to rewrite atof (?!)
eg
"-0.00801272 0.0320147-7.09668e-005 -0.0501062 -0.0879057 -0.101794"
* This program
* - explores recursively the (single Patient, single Study) directory
* - examines the ".txt" Dense multiframe files
* - Converts the files into 16 bits Dicom Files
* WARNING : directory must contain ONLY .txt files
(Probabely some more refinements to be added, but I wouldn't like to loose
the source)
//
// This program is used to convert huge amounts of functionnal (3D + T) MR images
// into 'image tiles' (one title per volume) processable by clinical softwares
//
Some old versions of VTK? ITK ? SeriesWriter weren't too much strict
when they did the job.
A lot of people has a lot of 'supposed to be Serie' directories
that, actually, aren't at all.
The following program converts them (hope so!)
'Acr-NEMA' data elements
0008,0001 Length to End
0008,0010 Recognition Code
0028,0005 Image Dimension
are now automaically pushed out by FileHelper::CheckMandatoryElements
when somebody rewrittes ACR into DICOM.
Thx to Mathieu for reporting the pb.
Port Mathieu's patch from gdcm 1.2 (in gdcmValEntry.cxx)
ENH: Better patch. Always pad with space except if VR is UI
(Shouldn't we think in unifying gdcm1.3 an cvs version -post gdcm1.3-
a lot of work is done twice; interesting enhencements are now in gdcm1.3
The only pb is New vs new, right?
jpr [Thu, 15 Jun 2006 14:22:33 +0000 (14:22 +0000)]
Little patch for strange PMS Gyroscan Intera images
( Item 'starter' has a tag : 0x3f3f,0x3f00, for no apparent reason )
'PMS' means 'Philips Medical Systems', as you can guess !
jpr [Wed, 31 May 2006 16:16:04 +0000 (16:16 +0000)]
Fix troubles in AnonymizeMultiPatient.
(It ensures (hope so) that all the Patient names will be different
after anonymization.
It will probabely replace AnonymizeNoLoad
jpr [Wed, 31 May 2006 16:09:29 +0000 (16:09 +0000)]
Add an open source code for MD5.
This will allowed to use a rough way to anonymize Patient's name, instead of
writing an arbitrary value.
(When we anonymized a multi-patient Directory, it leaded to a mees -all the
patients with the same name-)