Directly answering the question ...
In the past you haven't seemed interested in solutions involving composite columns and templates, but just in case I am wrong, here is one. It returns the width and height of the cover image. Performance isn't excellent because it must read the actual cover file off the disk. I suggest that if you use it then you display the results in book details and not in the book list making performance much less of an issue.
You would put this in a "Column built from other columns" with "Show as html" unchecked. You could check "Show in Tag browser" but it isn't obvious why you would want this.
Code:
python:
def evaluate(book, context):
from io import BytesIO
from PIL import Image
mi = context.db.new_api.get_metadata(book.id, get_cover=True, cover_as_data=True)
if mi.cover_data[1] is not None:
pm = Image.open(BytesIO(mi.cover_data[1]))
ans = f'w={str(pm.width)}, h={str(pm.height)}'
else:
ans = 'No cover'
return ans