Get a DICOM tag that is a sequence

[EstebanGuillen]

Is it possible to access a DICOM tag that references a sequence type using io.dicom.decode_dicom_data()?

Using pydicom I can acess take 0x0008,0x1155 like this:

ds = pydicom.dcmread(file)
reference = ds[0x0008,0x1115][0][0x0008,0x1140][0][0x0008,0x1155].value

How would I do this with the io.dicom.decode_dicom_data() function?

[yongtang]

@EstebanGuillen Do you have a sample dicom file I can take a look?

[EstebanGuillen]

Hello, Attached is a dicom that references another dicom file. When we annotate a dicom an additional dicom gets created; it contains a reference (pointer to another dicom file) and the annotation details (basically points for our use case). I need to get access to the reference and the points. I wasn't able to use 'io.dicom.decode_dicom_data' to access those values. Here is code using pydicom which can access the desired values. *import pydicomimport os file = './2.25.304589190180579357564631626197663875025.dcm' ds = pydicom.dcmread(file) #try to access the referencetry: path = os.path.dirname(file) reference = ds[0x0008,0x1115][0][0x0008,0x1140][0][0x0008,0x1155].value points = ds[0x0070,0x0001][0][0x0070,0x0009][0][0x0070,0x0021].value box = ds[0x0070,0x0001][0][0x0070,0x0009][0][0x0070,0x0022].value print('References the following dicom: ',reference) print('') print('List of tags:') print(ds) true_positive = path + '/' + reference + '.dcm' if int(points) < 5: print('Not enough points for a box', points) print(' box: ', box) except: #only annotation files will the a reference tag #a normal dicom with image data should through an exception pass* Thanks for your help, Esteban

…

On Sat, Feb 15, 2020 at 10:56 AM Yong Tang ***@***.***> wrote: @EstebanGuillen <https://github.com/EstebanGuillen> Do you have a sample dicom file I can take a look? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#779?email_source=notifications&email_token=ADYPWF5EECXDU3WCM5EZWKTRDAUDTA5CNFSM4KTSRKGKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEL3TCIY#issuecomment-586625315>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADYPWF6MVQR2SEFY5YPTG23RDAUDTANCNFSM4KTSRKGA> .

[yongtang]

@EstebanGuillen I am not able to see the attachment. Looks like GitHub filters out the attachment from the email. Can you attach the file to this issue page?

[EstebanGuillen]

@yongtang
Sure, here is the file:

example.zip

Let me know if this works

[yongtang]

@EstebanGuillen I didn't have a lot of experience with DICOM before, though I took a look at the dump in C++:

...
(0008,1115) SQ (Sequence with undefined length #=1)     # u/l, 1 ReferencedSeriesSequence
  (fffe,e000) na (Item with undefined length #=2)         # u/l, 1 Item
    (0008,1140) SQ (Sequence with undefined length #=1)     # u/l, 1 ReferencedImageSequence
      (fffe,e000) na (Item with undefined length #=2)         # u/l, 1 Item
        (0008,1150) UI =ComputedRadiographyImageStorage         #  26, 1 ReferencedSOPClassUID
        (0008,1155) UI [2.25.211904290918469145111906856660599393535] #  44, 1 ReferencedSOPInstanceUID
      (fffe,e00d) na (ItemDelimitationItem)                   #   0, 0 ItemDelimitationItem
    (fffe,e0dd) na (SequenceDelimitationItem)               #   0, 0 SequenceDelimitationItem
    (0020,000e) UI [2.25.34759593920092135343193448647498328319] #  44, 1 SeriesInstanceUID
  (fffe,e00d) na (ItemDelimitationItem)                   #   0, 0 ItemDelimitationItem
...

[yongtang]

@EstebanGuillen If I understand correctly, you are looking into accessing

[0x0008,0x1115][0][0x0008,0x1140][0][0x0008,0x1155]

which roughly equals to:

[sequence of 0x0008,0x1115] => 
    [item 0 of the sequence] => 
        [sequence of 0x0008,0x1140 in the item] => 
            [item 0 of the sequence] => 
                [value of 0x0008,0x1155 in the item]

[yongtang]

@EstebanGuillen The current decode_dicom_data does not handle sequences, and it does not handle recursive sequences as in your case.

The current decode_dicom_data only takes a uint32 which is the tuple of uint16 (e..g, 0x0008,0x1115) and convert into one uint32 as the input.

It is not very difficult to add the implementation for your need though.

I think a logical choice is to allow decode_dicom_data to take string as input. In this way, a string of [0x0008,0x1115][0][0x0008,0x1140][0][0x0008,0x1155] could be passed for accessing elements in (potentially recursive) sequence.

I might be occupied this week so may not work on this implementation until next week or later. However, I will mark this feature request as contribution/help welcomed. If anyone has an interest in get this done, please go ahead. I can point to the exact places to make this happen.

Otherwise I will come back later and see if I can add the implementation.

[yongtang]

@EstebanGuillen A PR #997 has been added to support getting DICOM tag that is a sequence, as was discussion in this thread.