This document summarizes Mark Kurtz's research on optical character recognition (OCR) through custom algorithms implementing techniques from computer vision. Kurtz created a program with some tested algorithms for OCR, documenting the results. He argues that current computer vision approaches are too specific and do not replicate the human visual system. Kurtz developed his research around the idea of using patterns, sequences, and predictions for object recognition, rather than exact template matching. Due to time constraints, he focused his implementation on OCR rather than full image classification and recognition.