Recognize text in screenshots

Question

Text recognising

Hey!
I need to recognize text in several thousand screenshots. The main problem is that this text is written in a small pixel font and consists of random letters (including those with diacritics), numbers, punctuation marks, special characters (which is why OCR programs like FineReader and Tesseract).

The fact that the text is clearly visible, and I also have a font with all these characters, makes it easier.

Do you know any ready-made solutions for recognizing such text? How would you solve such a problem yourself?

Thank you!

Answer 1

Scrap specifically for this purpose. A little old, but at one time no one better than him coped with this problem.
www.structurise.com/kleptomania/russian.shtml

Answer 2

For such purposes, I wrote my own recognizer. It worked primitively - it shredded the original image into strings, lines into characters, and recognized characters by comparing them with templates. About existence of ready tools for the similar did not hear.

Answer 3

You can try to drive it through OneNote, however, I can’t vouch for special characters.

Answer 4

FineReader seems to be teachable.

Answer 5

I agree with the FineReader learning option - try it. Recognized several hundred pages of typewritten and photocopied text in not the best condition as follows:
1. Divided the text into “chapters” depending on the deplorable state (torn, coffee, crumpled)
2. Run the first couple of sheets in learning mode (it took minutes 20)
3. Run recognition

Answer 6

But Abby fine reader and similar text recognition programs will not help. I recognized from pictures, though the usual large text and everything worked fine.

Answer 7

There is ABBYY Screenshot Reader, it is specially designed for such cases.