Why?
I needed an interactive tool that would assist (a team of monkeys) in recognizing a document structure from rasterized or scanned PDF forms because automated tools like Omnipage and Finereader offer no efficient method of form field data-binding, auto field naming, schema mapping and so forth. While the actual algorithms and the tool is closed-source for now, pXY.js was written to serve as the core.
tute: http://o-0.me/pXY/
repo: https://github.com/leeoniya/pXY.js