Hello @cscheeser ,
This is an image segmentation problem and not strictly a PDF-parsing problem. You can see here my response to something very similar:
Being able to do this kind of segmentation is actually state of the art, so it would require the advanced techniques shown in the link above.
1 Like