non-word recognition worsened/disimproved in tesseract v3.0.4 ?
84 views
Skip to first unread message
Jakob Kroeker
Jun 23, 2016, 5:56:03 AM6/23/16
to tesseract-ocr
Hello,
when trying to recognize non-words with mixed upper-lower-case letters
I observe issues with tesseract 3.0.4 while it is working with 3.0.2.
For example. try the following text (with Arial font) and see what you get:
Or is there a configuration for which it works in 3.0.4 as good as in 3.0.2?
Jakob
when trying to recognize non-words with mixed upper-lower-case letters
I observe issues with tesseract 3.0.4 while it is working with 3.0.2.
For example. try the following text (with Arial font) and see what you get:
FfVvZz
Zz
vZzOr is there a configuration for which it works in 3.0.4 as good as in 3.0.2?
Jakob
Reply all
Reply to author
Forward
0 new messages