[tesseract-ocr] Superscript Broken

2015-01-31 Thread Jim
OCR the attached image in PSM_SINGLE_BLOCK using tesseract 3.04 latest with default eng.traineddata and it outputs 2\nx instead of x2. Scanning a document with a bunch of superscripts like this makes a mess.Does anyone know any options to affect new line creation without changing PSM or do you

Re: [tesseract-ocr] Help needed in understanding source. New to tesseract.

2015-01-31 Thread Allistair
If you start by learning C++ then you will realise the entry point to a C++ program (main), you can then trace the various calls either manually, methodically stepping through files and functions by searching. You could potentially look at tools that allow call stack debugging/tracing. I am not