Since iOS 11, Apple released Vision, their powerful computer vision framework, including OCR capabilities. I strongly recommend using it because it's very easy to use and well performing. That said I'm no longer maintaining this project, I'm archiving it and leaving it here for history.
This iOS app uses gali8/Tesseract-OCR-iOS to recognize text from images taken from live camera preview.
This project uses Cocoapods as a dependency manager, after checkout, run pod install
from inside the project folder.
If you don't have Cocoapods installed, here's the official reference to get you started.
Recognition is now coded to work with Italian and English language. In order to use another language
- Checkout tesseract-ocr/langdata
- Replace all
ita.*files in theTextReader/tessdatafolder with the desired ones (all of the files in your language folder are needed if you don't change recognition mode) - In
ViewController.swift, row 17, replaceitawith the 3 letters code of the language you chose - Edit the char whitelist in
ViewController.swift, row 26, to tailor alphabet and punctuation to the ones you expect to find in your text, or delete the row to keep it general.
As of my experience I suggest to maintain eng.* files even if you don't actually need them because they empirically seem to
enhance the precision of the recognition of other languages as well.
The app, as is, uses both recognition modes: Tesseract and Cube.
By using both of them it's way more precise but slower too, so you may want to try using Tesseract only mode first, that's the
fastest, to check whether it's enough to suit your needs. In order to change it, during Tesseract initialization
(ViewController.swift, rows 25-27) set G8Tesseract.engineMode to .tesseractOnly.
If you find that's good for you, then you can delete all of the files in the TextReader/tessdata except for the ones called
*.traineddata, because the other files are used only by Cube. In this way you'll sensibly reduce your app size, too.
For more configuration fine tuning, I strongly advice you to check these pages of the gali8/Tesseract-OCR-iOS documentation