It also converted the line numbers in the document, such as pdf. I think this should be improved, ignoring the line numbers or watermarks.