Processing diffusion images from Philips scanners

[Yann Cointepas]

During HBM 2005 in toronto, someone gave me a 32 directions diffusion acquisition from a Philips scanner (thank you Christine). There are several points that have to be handled before BrainVISA can automaticaly import and process these data:

• DICOM format: our current DICOM reader is unable to properly handle the files I received. We already started to improve DICOM management but this improvement will not be available before a few weeks(at least). I have been able to convert the data using medcon (http://xmedcon.sourceforge.net/) to extract raw data in a *.ima file and manualy creating a *.dim header.
  Interlaced slices: Unlike other systems, Philips diffusion slices are not ordered to keep volume continuity. Slices of different volumes are interlaced (first slice of T2 volume, first slice of first direction volume, first slice of second direction volume, etc.). This structure can be handled in BrainVISA by first reordering the slices but it would require to read input image several time. We will try to avoid this by changing the low-level tools used to import image. This will not be done before the new DICOM reader is finished.

I will make a temporary experimental BrainVISA import process (with low support) working with the current release. This process will use medcon (one who wants to use this process will have to install medcon) and will probably only accept DICOM files. Let me know if you are interested by this process.

[Manik]

Hi,

if you do such a process, could you make one that can be used to import any dicom image as a GIS image ? Because I have some DICOM images (3D-MOTSA sequences) that can be imported with medcon but not by BrainVisa (I gave you one of those a few month ago).