I am looking for a software/library (preferably open-source) to read bitmaps containing texts in the OCR-B font, which is used e.g. in passports machine readable zone. I have tried tesseract, but it gives me poor results. I assume that this has to be a solved problem, when OCR-B is international standard so I'd think that such software is is not exclusive to airport staff, but I am unable to find anything.
Have you tired OCRSDK.com from ABBYY? It is not open-source, but has some amount of free usage for beginners. Very high accuracy. Here you can play around with your own images without any registration and see what it is capable of.
Disclaimer: I work for ABBYY.
