fix color error

[aporter-turing]

This PR fixes the convert legacy error related to adding RGB dicoms

[CPBridge]

Thanks @aporter-turing for reporting this and for your suggested fix. Unfortunately while your code may "work" and successfully produce a DICOM object, my reading of the standard is that that object would not be a valid Legacy Converted Enhanced MR Image.

There is a constraint that the PhotometricInterpretation have the value "MONOCHROME2", would would appear to rule out any color images being included in the Legacy Converted Enhanced MR Image altogether (source)

The Photometric Interpretation (0028,0004) defined in Section C.8.15.2 is MONOCHROME2.

So instead, we should probably check this constraint and raise a more intuitive error message if any input image is not "MONOCHROME2".

[aporter-turing]

Hi @CPBridge

Thank you for the prompt response. I noticed the link you sent was for a CT image. When I look at the MR details here I don't see the same restrictions for PhotometricInterpretation, would it be possible to allow for this conversion for MR only data?

[CPBridge]

Oops yes the link was wrong but I'm afraid the same is true for legacy converted MR (source).

If you want a multiframe MR image, I thenk you would need to use the Enhancnd MR Color Image, but highdicom does not support this at the moment (maybe someday)

[CPBridge]

Closing in favour of #321