Some notes to self on document layout analysis. I’m revisiting the problem of taking a PDF or a scanned document and determining its structure (for example, where is the title, abstract, bibliography, where are the figures and their captions, etc.). There are lots of papers on this topic, and lots of tools.