pix2tex - LaTeX OCR

The goal of this project is to create a learning based system that takes an image of a math formula and returns corresponding LaTeX code

Data

We need paired data for the network to learn. Luckly there is a lot of LaTeX code on the internet, e.g. wikipedia, arXiv. We also use the formulae from the im2latex-170k dataset.

Fonts

• Latin Modern Math
• GFSNeohellenicMath.otf
• Asana Math
• XITS Math
• Cambria Math

Requirements

Dataset

In order to render the math in many different fonts we use XeLaTeX, generate a PDF and finally convert it to a PNG. For the last step we need to use some third party tools:

• XeLaTeX
• ImageMagick with Ghostscript.
• Node.js to run KaTeX
• de-macro >= 1.4
• Python 3.7+ & dependencies

Contribution

Contributions of any kind are welcome.

Acknowledgement

Code taken and modified from im2markup, arxiv_leaks