Tesseract problem with mftraining stage

Question

I successfully created a window file with tesseract
now after running unicharset_extractor by
creating a unicharset file that looks like this:

...
n 3 NULL -1
s 3 NULL 23
t 3 NULL 43
...

I continued this team

mftraining -U unicharset -O testlang.unicharset testlang.tr

to get the following error

Reading testlang.tr ...
testlang has no defined properties.

Error: Illegal short name for a feature!

Answer 1

I have never worked with Tesseract, but it looks like there is an open problem in the error database that is very similar to your problem: http://code.google.com/p/tesseract-ocr/issues/detail?id=385

This seems to be due to the fact that scientific notation is not properly supported by some functions.

, . , , .