How to recognize characters in an image using OCR without specifying any language?

Question

Is there a way to recognize characters in an image using OCR without specifying any (human) languages?

Note: I have a set of characters (with a specific font) in an image. The content that consists of these characters in meaningless (from a language point of view) like a car no.

Is there a way to (programmatically or by using a product; open-source is preferred) recognize them?

Answer 1

Have you considered using Tesseract-OCR? As of version 3, tesseract can work well with left-to-right languages. You can also train tesseract.

I have not used it myself but if you have a limited set of characters you want to recognize -- from different languages, maybe you can modify the training procedure to suit your needs.