






Sorry for late answer: your answer has worked for the box.
I have tested your last version, it is clearly better and very interesting.
I have tesseract 3.01 and 3.02 on my computer
I have already succeeded to make a trainneddata but I had to change between the both version:
for the box I have only good result with 3.01 with 3.02 I have only empty file.
with 3.02 I have Tesseract Open Source OCR Engine v3.02 with Leptonica
Empty page!!
Empty page!!
and tst.mise.exp1790.box an empty file...
What does RTL mode mean?
When we create a box, does the language is useful?
When we train a file why not open it directly when it is done?
The possibility to same the image file is usefull when i use photoshop to make the picture the file can be read by your software but not by tesseract. After saving it inside I have a correct version of my tif for tesseract.
for other files i need to use tesseract 3.02 3.01 does not work for all
Maybe it could be interesting to have the correct version files of tesseract training adapted to your software directly include with it. Or a precise version (url) of the right version to install
What is StdErr mode? in Dos output result it could be interesting to have the terminal line used.
Actually I have a traindatta file but when I use it, tesseract never end the recognition. I have to Ctrl +C
It could be nice if last folder used remains in memory (the setting option that you added is a nice thing)
Thanks for the new version: it is a very nice evolution :)
Barthelemy
Ps: I have seen some other problems but I will send it in new messages when i will see it again to give you the context