How to reliably recognize old texts under Linux from the command line?

Question

System administration

Like, for example, this text . There, by the way, there is a layer of recognition text, but you can see how this recognition is ugly, and you need a search by last name.

Answer 1

It was a long time ago - it was necessary to recognize Belarusian texts (symbols i, ў in Cyrillic texts), and FineReader did not know how to do this then, but it has training. Conducted training on 2 pages, he recognized the rest without any problems.
So something like this: take a FineReader or another recognizer that is available for Linux, train it manually, and then set the trained recognizer on scans from the command line.