Tesseract OCR (3.0.2) on iOS (6.1++, no arm6),
Just putting together the bits and bytes to include OCR in an iOS application. Readonly mode, no training.
This project is by no means targetted to obtain a commercial product but should serve as a jump board to create your own OCR aware iOS app.
It contains a classic Master Detail Controller with Core-Data, SenUnitTest
Also note that the analysis on the phone is slow for bigger images.
This work is based on Tesseract http://code.google.com/p/tesseract-ocr/ and Leptonic http://leptonica.com/ for some of the image processing parts.
See also install_newest.txt to understand what to do if you want to compile the 2 libraries on your own (it is not necessarily a trivial task and might anytime break as it envolves a lot of stuff to work together). You might also want to do this to streamline the libs.
Run unittest for regression testing, the results are predictable so TDD is a YES here. But we are dealing with a particualr kind of test as it also affects the quality and not only a boolean win/fail. Some tests will fail in this version (that depends on the traineddata and the fonts used in the image). This is not necessarily a bad thing.
if you call many times the analyseImage method you keep enqueueing calculations, there is no "empty-queue-mechanism" implemented (up to you).