What to use to create a checkmark/box image recognition algorithm?

I'm not asking on how to do something, but simply, I need direction. I'm not well versed in this field, so I do not know where to start.

Anyways, my python algorithm should take in a pdf and extract the image out of it. That image contains a list of questions and typically three empty checkboxes and one ticked checkbox for each question. My intention is to link each correct checkbox to the corresponding question. For example: On the image we have:

Question 1. A:(empty) B:(empty) C:(ticked) D:(empty)

Question 2. A:(ticked) B:(empty) C:(empty) D:(empty)

So, my algorithm should return an array resembling ['C','A'].

Does anybody have any idea what library I should use for python, or do you have any tips or hints?

Thank you and I appreciate it a lot.