Hello,
I'll be glad to help you. I'm doing custom OCR implementations for more than 4 years and I'm quite season in image processing for OCR. Usually, I work with a highly modified Tesseract engine or with Transym TOCR. Not all the images that your eyes can "read" can be processed - if you can acquire the images yourself, meaning that you can control the raw data quality, that's fine, otherwise you have to rely on enhancing algorithms for binarization, filtering and so on.
My OCR solutions are desktop-based or web-based (WCF - SOAP or RESTful), by request or by problem to be solved. I'm developing only for Microsoft OS with VS2012.
I'm looking forward for some samples.
Regards,
Dragos I. Stanciu