Keeping correspondence between DICOM tags

156 views
Skip to first unread message

Razvan Bocu

unread,
Mar 21, 2017, 5:10:18 AM3/21/17
to RSNA MIRC CTP/TFS User Group
Hello,

There is a special use case that I should configure the anonymizer for. Thus, in order to be visualized in a particular application, the anonymized DICOM files maintain the association between them considering the same value for SOP###UID tags. Thus, as an example, the values for SOPClassUID and SOPInstanceUID is identical. Nevertheless, it should stay identical in the anonymized form, as well. What kind of anonymization setting should I specify in the script so that identical tag values stay the same after anonymization?

John Perry

unread,
Mar 21, 2017, 8:59:56 AM3/21/17
to rsnas-ctpmir...@googlegroups.com
You can use this script for all the UID elements that you want to remap:
 
    @hashuid(@UIDROOT,this)
 
To do it correctly, you should get a UID from some official source and set the UIDROOT parameter in the script to that UID.
 
Note: You should replace the value of the UIDROOT parameter in the anonymizer script file with a UID that you own. The value of the UIDROOT parameter in the default DicomAnonymizer script is set to "1.2.840.113654.2.70.1". I don't have any memory of how that came to be; it used to be "9999" as a convenience for testing. In any case, you should change it.
 
Note: Some UIDs – for example, TransferSyntaxUID and SOPClassUID –  should not be remapped because they refer to specific things in the DICOM standard.
 
With the script above, two identical UIDs will be mapped to identical values, whether they occur in the same data object or in different ones.
 
JP
--
You received this message because you are subscribed to the Google Groups "RSNA MIRC CTP/TFS User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rsnas-ctpmirc-user...@googlegroups.com.
To post to this group, send email to rsnas-ctpmir...@googlegroups.com.
Visit this group at https://groups.google.com/group/rsnas-ctpmirc-user-group.
For more options, visit https://groups.google.com/d/optout.

Razvan Bocu

unread,
Mar 22, 2017, 7:02:01 AM3/22/17
to RSNA MIRC CTP/TFS User Group
Hello,

I tried to apply the function that you suggested to two DICOM files. I checked that the original, not anonymized files, have the same SOPInstanceUID in their header. Unfortunately, the anonymized files receive different values for the SOPInstanceUID, and they should stay the same in their anonymized(hashed) form, otherwise they are unusable by a certain software package.

What could I be doing wrong, how could I anonymized and get the same anonymzied values for tags that are identical in the original DICOM files?

John Perry

unread,
Mar 22, 2017, 7:29:52 AM3/22/17
to rsnas-ctpmir...@googlegroups.com
Please send me your config.xml file and your anonymizer script file.
 
If you can, please also send me one image that exhibits the problem.
 
Send them to my personal email (john...@dls.net).
 
JP
 
Sent: Wednesday, March 22, 2017 6:02 AM
--

Razvan Bocu

unread,
Mar 22, 2017, 8:16:34 AM3/22/17
to RSNA MIRC CTP/TFS User Group
I sent the requested email to the specified address.


On Tuesday, March 21, 2017 at 11:10:18 AM UTC+2, Razvan Bocu wrote:

Brian O'Brien

unread,
Feb 15, 2023, 6:02:23 PM2/15/23
to RSNA MIRC CTP/TFS User Group
I have pretty much the same need.
I need all VRs of type UI (Not SOP Class UID, Transfer Syntax.. and i am not sure what else also is a VR of type UI)
I believe the method is to create a table that take a study UID, searches for it in a dictionary and if it is found returns a list of UIDs that this UID has been 'translated' to.
Each time an anonymization is done this new remapped UID gets added to the Value of the dictionary of translated UIDs.
I say EVERY UID because I want to recurs the datasets and sequences for Referenced UIDs and those referenced UIDs must be remapped to maintain referential integrity.

If anyone has an anonymizer xml or script that does this I think the whole community would benefit from this pipeline.
Reply all
Reply to author
Forward
0 new messages