I have many images containing a font of known size. Tesseract does well when it finds the correct bounding boxes of characters in this font, but it often merges characters. I'd like to tell it what the character size is, since I can determine this accurately myself.