Dear all,

DICOM PS3.16 2023b specify Context Groups to describe methods used by an application under CID 7050 De-identification Method (*).

Following those, there are cases where an application may convert a DICOM instance into an alternate instance using a different Specific Character Set (eg ISO_IR 100 to ISO_IR 192). This operation may occurs more often as Web application typically expect JSON encoded as UTF-8.

The issue is that this operation is not documented and a receiving application does not know the history of the SpecificCharacterSet, since the previous value has been overriden.

There are obvisouly cases where this conversion process may causes some confusion:
1. Typically a latin1 encoded value located directly inside a UN attribute (either IVRLE or EVRLE+UN), or indirectly located inside a UN sequence (typically IVRLE or CP-246 sequence). The sending application will obvisouly not convert this attribute, however the receiving application may. See for example (**).
2. Sometimes it is not always possible to convert directly into UTF-8 and one either needs to "transliterate" characters, or simply discard those problematic characters. A receiving application may indicate that characters were modified in the process.

I think it would make sense to define a set of Context Groups for those, so as to hint a receiving application about the history of the possible SpecificCharacterSet conversion operations/values.

Comments ?

(*) https://dicom.nema.org/medical/dicom/current/output/chtml/part16/sect_CID_7050.html
(**) https://support.dcmtk.org/redmine/issues/1073