Missing

From gdcm

TODO

See actual TODO file

Things missing from GDCM implementation

• Major
  • Conversion from DICOM to XML
  • a gdcm::Parser implementation (aka SAX)
  • Command/Observer implementation
  • multi-frame related pb
    • A 'frame by frame' reading, even for compressed images
    • single frame extraction from a multi-frame DICOM file
  • Big Endian machine are not tested, likely to crash or produce incorrect results. If you own such machine and would like to contribute a GDCM dashboard, please drop a note in the GDCM mailing list.
  • Enhanced MR/CT Image Storage are poorly supported (only simple cases will work).
  • deflated syntax should be padded with a \0 when appropriate
• Minor
  • GDCMPolyDataReader and 12 ECG data.
  • gdcm::Scanner & SkipTags
  • gdcmdump -p and color on WIN32 terminal, make the color a compile time option
  • Modality LUT: Tag(0x0028,0x3000) is not taken into account
  • Fiducials Object
  • A Validator (PS 3.3) aka dciodvfy
  • gdcm.PythonDataSet and gdcm.PythonTagToValue are a hack to get around annoying swig issue, need to be fixed !
• Feature request
  • Papyrus
  • FFMPEG plugin
  • MPEG2
  • Ruby wrapping
  • Replace swig wrapping with Boost.Python (http://lists.boost.org/boost-users/2008/01/32997.php)
  • Display (all or part) of UN or OB Data Elements that contain both printable and non printable char, in a human readable way (ala hexedit)
  • Virtual Big Endian is not fully supported (DataSet encoded in "implicit Value Representation" and big endian). (MM: I do not think this is even remotely possible, think IVRLE SQ / cp 246)

Done

• Memory-efficient bridge between GDCM and VTK, ideally in the non-streaming case just as much memory required as the size of one slice of data (in the case of 2D series).
  • 2D Series only require an extra 'line' overhead instead of the whole image.
• A gdcm::SeriesSorter and a gdcm::VolumeSorter are currently missing, they would allow the sorting of filenames according to user criteria
  • See gdcm::Sorter & gdcm::IPPSorter
• An acces to encapsulated images (within an 'Icon Image Sequence', inside a Dicom file, or inside a DICOMDIR)
  • Update:There is now a gdcm::IconImage
    • gdcm::IconImage should derive from gdcm::Image to allow processing of RGB/PALETTE (reuse code)
      • Fixed
• Curves
  • See gdcm::Curve
• vtkGDCMImageWriter should implement SetFileDimensionality
• Modality LUT: linear rescale should be applied (http://visualisation.tudelft.nl/~cpbotha/thingies/IM-0001-0001.dcm).
• Fix the Origin in vtkGDCMImageReader, depends on Image Orientation (Patient). Finish up gdcm::Orientation integration.
• Add FileLowerLeft to vtkGDCMPolyDataReader -> IMPOSSIBLE
• Conversion back/forth from raw to jpeg compressed
• Segmented Palette
• RT Object
• Conversion from raw file like jpeg into DICOM
• 16 Bits LUT table are not supported
• JPEG-LS
• Icon Image SQ can be compressed and/or PALETTE COLOR
• ITK bridge is currently very poor
• #Net wrapping
• DICOMDIR