search

icometrix / dicom2nifti

MIT License
329 stars 78 forks source link

Converting Rgb RCBV images #5

Open varghesealex90 opened 7 years ago

varghesealex90 commented 7 years ago

Hi;

I was wondering if it was possible to convert Relative blow volume maps (RGB) to nifti using the software you have created

abrys commented 7 years ago

Hello,

Currently we do not support any RGB data. An example dataset (anonymous) would be helpful to implement support somewhere in the future. Also we I am wondering how the RGB data should be stored. Using the 4th dimension?

neurolabusc commented 7 years ago

@abrys I am providing a sample color DICOM image here. It was created during the installation of our Siemens Prisma on D13D using the vendor sequence. It is a nice example because of the small file size. Modern multi-band sequences would generate better but larger images. Further, I usually set up our sequences to NOT create color maps: 8-bit images have low precision and one gets much more accurate diffusion maps after off-line undistortion using Eddy/Topup. One final caveat: this dataset exhibits a common bug in dervied Siemens RGB images in that it does not include the spatial transformation matrix that was included in the raw data. I believe this has been fixed in more recent Siemens software. It is my brain and I am releasing this image to the public domain https://www.dropbox.com/s/5yauu814gnveo8t/color_dicom.zip?dl=0 You can test your output versus dcm2niix and view the output with MRIcroGL. You would set the NIfTI datatype to DT_RGB24 (=128) and save the data as packed triplets (RGBRGBRGB...). Note my sample will look odd due to the lack of a spatial transform. Further, many NIfTI tools will not read RGB (e.g. SPM) and that some older tools will have problems displaying these images (the Analyze format that preceded NIfTI assumed datatype 128 was planar, so for each slice in the volume the data was RRR...RGGGG...G.BBB...B).

neurolabusc commented 7 years ago

The example I provided is 0028,006 = 0 (data saved as RGBRGB...) you may also want to handle Planar data (0028,006 = 1, data saved as RRRR...RGGGG...GBBB...B). My hardware does not generate these, but there is an example here http://www.barre.nom.fr/medical/samples/ and an explanation here: http://dicom.nema.org/medical/dicom/2014c/output/chtml/part03/sect_C.7.6.3.html#sect_C.7.6.3.1.3

varghesealex90 commented 7 years ago

Hello all;

I can share with you images acquired from Siemens avanto (1.5 T) and GE (3T) you. It would be great Arne, if you could share your personal mail id with me.

Regards Varghese

On 26 July 2017 at 17:38, Chris Rorden notifications@github.com wrote:

The example I provided is 0028,006 = 0 (data saved as RGBRGB...) you may also want to handle Planar data (0028,006 = 1, data saved as RRRR...RGGGG...GBBB...B). My hardware does not generate these, but there is an example here http://www.barre.nom.fr/medical/samples/ and an explanation here: http://dicom.nema.org/medical/dicom/2014c/output/chtml/ part03/sect_C.7.6.3.html#sect_C.7.6.3.1.3

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/icometrix/dicom2nifti/issues/5#issuecomment-318034251, or mute the thread https://github.com/notifications/unsubscribe-auth/AJqPzn73T4wVQgRgRzu-fU5Xg3JuBGKsks5sRyxTgaJpZM4OVTCL .