This update fixes the confidence value issue I reported in #56
fixed the memory leak
Modified characterBoxes function to return characters in order
Removed unused lib files
Added some null checks to fix a bug where no text is recognized
Note: Building the Tesseract OCR Aggregate product will yield a warning about how the Tesseract and Leptonica lib files don't work for the x86_64 architecture. However, everything still works in the simulator for iPhone 5 and lower, as well as on all physical devices. I guess because the libraries work okay with the i386 target.