wordacc does not count words that consisted of numbers.

[GoogleCodeExporter]

What steps will reproduce the problem?
1. I compared OCR output of a receipt image with ground truth using wordacc.
2. Ground truth has this line: PLASTIK-KARTON BARDA %18 *1,75
3. OCR output has this line:   PLASTIK-KARTON BARDA %18

What is the expected output? What do you see instead?
Although output is different with the ground truth, wordacc gives 100% 
accuracy. I thinki that *1,75 (price) is important and I have to count a price 
as a word for my study.

What version of the product are you using? On what operating system?
R2 on Ubuntu 14.10

Please provide any additional information below.

Output (4-results.txt) and ground truth (gt.txt) are attached. Wordacc gives 
%100 accuracy, but they are not equal.

Original issue reported on code.google.com by sozars...@gmail.com on 13 Jul 2014 at 9:10

Attachments:

• gt.txt
• result-4.txt