Extracted image is not transparant

[Gardener12]

Hello, dear Github friends,

I have a question about PyMuPDF.
I have created a pdf (test.pdf) with only one transparent image in it (see 'original image.png').
When I extract the image with PyMuPDF, the result it's not transparant (see 'pic-11.png').
Also I notice, the original image has 32 bit depth, the extracted image has 24 bit depth.

What can be the cause/solution?

I'm new here on Github, so my apalogies for any mistakes I make here.
Thank you for your time!

p.s. I uploaded the python code as extract.txt, because otherwise the code was not displayed well.
p.s. I use Python 3.9 and MyMuPDF is updated
p.s. Whether I make the pdf in RGB modus or CMYK, the result is the same

--- the output about the image ---
{'number': 0, 'bbox': (36.0, 46.0, 173.25, 142.0), 'transform': (137.25, 0.0, -0.0, 96.0, 36.0, 46.0), 'width': 183, 'height': 128, 'colorspace': 1, 'cs-name': 'Indexed(1,DeviceRGB)', 'xres': 96, 'yres': 96, 'bpc': 8, 'size': 1641, 'digest': b'\xbf\xe52<\x96\xa8\x87x\xd7Y\xf4\xdc7fg\xaf', 'xref': 11}

{'ext': 'png', 'smask': 12, 'width': 183, 'height': 128, 'colorspace': 1, 'bpc': 8, 'xres': 96, 'yres': 96, 'cs-name': 'Indexed(1,DeviceRGB)', 'image': b'.....'}

test.pdf
extract.txt

[JorjMcKie]

If there is an smask > 0, then this xref contains the alpha channel (so to say) for the base image and you must make a composite pixmap of the both to recover the original image ... along this snippet:

base = fitz.Pixmap(doc. xref)
mask = fitz.Pixmap/doc. smask)
pix = fitz.Pixmap(base, mask)

[Gardener12]

Thank you, Jorj, for your fast reply! I will try it!

[JorjMcKie]

A peculiarity of PDF is that it stores transparent images under separate xrefs. So one has to collect those pieces again for making the original.
You can achieve the same result using MuPDF's CLI like that: mutool extract -a test.pdf 11.

[Gardener12]

Thanks for this information, Jorj. I didn't know that about pdf's