Hello,
I am trying to train tesseract using existing images, 'makebox' works well but 'box.train' results with "Error during processing.", nothing more, nothing less.
drich@home$ tesseract img.tif eng.captfont.exp0 nobatch box.train
Tesseract Open Source OCR Engine v3.05.00dev with Leptonica
Error during processing.
Before using the git-repository version, I was using one from ubuntu repos which was resulting in a segmentation fault in PAGE_RES_IT::start_page(bool)
Here is one of the images I am using (copy-paste content to browser's address bar) :
http://pastebin.com/mkzWJi9E