DICOM @ OFFIS

Hi, I'm quite puzzled as to what causes a VR to change from UT to unknown, after sending an image file to a PACS. Here's the output of the file before I send the file (using dcmtk obviously): W: DcmMetaInfo: No Group Length available in Meta Information Header # Dicom-File-Format # Dicom-Meta-Inform...

OK never mind, I just noticed the setEos() API which is apparently needed.

Hi, I'm trying to serialize a DcmDataset for interprocess communication, without going to a file, and am not having success. I trimmed down my code to this, and even this doesn't work: BYTE buffer[4096]; DcmOutputBufferStream outStream(buffer, 4096); E_TransferSyntax xfer = EXS_LittleEndianExplicit;...

The only fail-safe approach IMO to compute the exact spacing between images is to use the ImagePositionPatient (0020,0032) combined with the ImageOrientationPatient (0020,0037) information, and if not present, well you're out of luck.

How to estimate the size of dataset before writing to local drive ?

AFAIK, you can use DcmDataset::getLength() to know the size of a dataset.

Not sure I understand, I want to extract from a DcmFileFormat object only the tag info, no pixel data nor overlay data, etc. Right now it's something like: DcmFileFormat ff; ff.loadFile(blahblah); DcmDataset* ds = ff.getDataset(); ds->remove(DCM_PixelData); Not sure it's the best way though. There i...

Hi,

what's the _best_ way to obtain only the tag information (with meta-info) using dcmtk? I'm looking to have a DcmDataset object (I presume) with no pixel data, thus having only the various tags.

Thanks,

Norm

OK great it's all clear now. As I said, it's all implemented and working, and I don't really have the need to feed the data dictionary, so I just left it off.

Thanks for your patience

But this is the definition of PRIVATE_ELEMENT1_TAG
seems to contradict your statements above, as there is a group.

I apologize for being so dense, I'm really trying to understand here.

yes that's exactly how I understand it. So in my case, I use group 0x0121, so that doesn't vary. It's all implemented already in my app, but most of the confusion on my part was with the need (or not) to add elements to the data dictionary. I still don't see how that can work, because of possible co...

Hmm, I had already read Part 5 section 7.8, and have just re-read it again, and I understand exactly the opposite, i.e. group (gggg) is the one thing that doesn't vary, and one is supposed to try inserting gggg,0010 to gggg,00FF until one succeeds, and when successful it means that the element inser...

er,
that is.

I'm running in circles here. I guess what I'm getting at is this: - What's the use of adding your private tags to the dictionary if you're not sure of the numbering you'll ultimately use because of possible conflicts? So to check for a Private Creator Tag I have to do something like this(assume that...

That was all pretty clear already, but I don't see in the above HOWTO how to resolve collisions with another vendor.

Boy this private tag business sure is confusing. I have questions: - Do we need to add entries to the data dictionary? If so, which ones? The private creator tag or the data element tag? I don't want to recompile, so I'm trying to add dynamically. - How does one handle the possible numbering clash w...