There are a number of OCR tools for regular text. Most of those tools rely on regular text's tendency to put all of the characters on a single baseline. However, math uses a complicated system of many nested baselines, and characters like radical signs which do not sit in sequence with other characters, but around them. All of these differences mean that traditional OCR does not immediately work for math. The purpose of this project was to create an OCR system that worked with images of printed math.