I am looking to use OCR to gather text from a specific location in an image, of known size, colour, baseline etc. I know Tesseract is very clever at finding text of all sizes and orientations, but I suspect this significantly slows the algorithm. Given a lot of information being locked down, including the expected characters, by what means can the OCR solution be most simplistically found?