Make the program quicker

The program runs a lot of titles in Tesseract, which seems to be quite time consuming.
Is it possible to minimise the number of unnecessary runs?

- [ ] A reading of a single card should take no longer than a second in release mode (not using the fanciest computer).

No source files or tests are named in the issue. Start by profiling the Tesseract calls involved in reading one card and measuring release-mode performance on a modest computer; done means unnecessary runs are reduced and a single card is read within one second.

Written by the indexing model from the issue text.