Closed chrisvanrun closed 2 years ago
Don't they need to be added to https://github.com/DIAGNijmegen/rse-panimg/blob/d95ab0f4ae2bf728442859b1534c1291a4d47540/panimg/models.py#L103 rather than here? Doesn't this put us in the strange situation where MHA generated from DICOM by panimg would then be rejected by panimg?
Also something similar to this was added in https://github.com/DIAGNijmegen/rse-panimg/commit/4f6ac2bff46e3101e26ee1731a3d5e8659cda1b2 but then we were told to remove it https://github.com/comic/grand-challenge.org/issues/1464#issuecomment-674558433 and https://github.com/DIAGNijmegen/rse-panimg/commit/87f04b54597401449968c127be5a22a337cc530f, unsure why this has now changed.
I agree with the philosophy that any image content / interpretation changing should be pre-applied to the data therein before trying to mimic the DICOM tag.
It does come with the cost of having to add custom computational code to handle tags like photometric interpretation.
See title.
Related to: