Problem writing dicom file after decompression

[jstefanis]

Hello, I have a multiframe dicom archive of CEUS and B-MODE images. After the update to version 2.x of pydicom lib, the images started to appear green.
When searching the forums, I managed to find out the problem, related with Photometric Interpretation and I was able to fix it and convert the sequence to RGB for further analysis.

My problem now is that I can't write the processed RGB sequence to a new dicom file.
I change the Photometric Interpretation and try to save, but error messages appear
like "AttributeError: Encapsulation required for compressed transfer syntaxes",
"ValueError: (7FE0,0010) Pixel Data has an undefined length indicating that it's compressed, but the data isn't encapsulated as required. See pydicom.encaps.encapsulate() for more information"

I'm stuck because I'm not experienced enough to resolve the problem and the documentation doesn't help me a lot. Any further help would be welcomed.
The dicom archive info is attached below.

Kind regards.

ceus.log

[jstefanis]

a code sample:

#ceus_fixed is the loaded dicom dataset

result = pydicom.pixel_data_handlers.util.convert_color_space(ceus_fixed.pixel_array, initialInterpretation, targetInterpretation)

ceus_fixed.PixelData = result.tobytes()
ceus_fixed.fix_meta_info()
ceus_fixed.save_as(filepath, write_like_original=False)

[mrbean-bremen]

Your data is in a compressed transfer syntax (JPEG Baseline), while pixel_array is uncompressed data (it does the decompression on the fly). So, either you have to change the transfer syntax to match your data (e.g. use Little Endian Explicit), or you have indeed to compress and encapsulate the data first. It is probably easier to write the data uncompressed by changing the TransferSyntaxUID, even if the file gets larger that way:

ceus_fixed.PixelData = result.tobytes()
ceus_fixed.fix_meta_info()
ceus_fixed.file_meta.TransferSyntaxUID = ExplicitVRLittleEndian
ceus_fixed.save_as(filepath, write_like_original=False)

[jstefanis]

Thank you very much about your help. It works like a charm.
I would like to ask if you don't mind, in the case that I want to re encode the frame set in its default compression (baseline JPEG), which are my concerns and the steps I should follow.
Kind regards

[mrbean-bremen]

I have to admit that I'm not experienced with compression either. As you can see in the documentation, pydicom itself doesn't support compression besides RLE. The decompression is done using thirdparty libraries (Pillow, GDCM, pylibjpeg) via a pydicom interface, but there is no interface for compressing data via thirdparty libs in pydicom yet. @scaramallion may have more insight here...

[scaramallion]

Sorry, I can't provide specific advice because I haven't confirmed that Pillow or GDCM will encode to JPEG Baseline in a way that meets the requirements of the DICOM Standard (i.e. subsampling, using the correct JPEG process, encoded color space, etc).

That said, it's likely that using Pillow will at least give you a JPEG that can be decoded by other programs, and the Photometric Interpretation will most likely be in YCbCr if you pass RGB data to Pillow.