OCR (Single) seems off.
- This topic has 1 reply, 2 voices, and was last updated 7 years, 3 months ago by
Kirill Bulatnikov.
Kirill Bulatnikov.-
Posts
-
David Yoshida
The OCR (Single) property appears to shift the selection several dozen pixels and fails some of the compares for existence.
Most of the compares are passing with 100% precision, but I don’t see how.When I change the property to be image, the check appears to be more accurate, at least the logging does.
Am I missing something? Thanks.
-David Yoshida
Kirill Bulatnikov
Hello David, thanks for using ZAPTEST
The object been recognized (you can see 100% precision) while not reflecting correctly in the report. Are you using IE to open test reports from ZAPTEST? I have seen before cases of difficulties with expand of the results in IE due some permission rules etc. And opening of the test results in Chrome should probably resolve the problem.
As for OCR vs Image – you do not need to use image recognition type for the text, OCR is proven engine and works good in 99% cases. The reason why you may have negative result with OCR is that the object you are trying to validate is
– out of visible area – in this case you may scroll to it using “ScrollTo” method
– relatively small/low resolution and cannot be rendered by OCR with enough quality, in this case we do not recommend to use it because of not sure in perfect recognition
– having different text color then black and different background then white – try to play with Text Color/Color Tolerance/Gamma settings of OCR objectHope that helps.
Thanks,
ZAPTEST Team.
David Yoshida
Kirill Bulatnikov- You must be logged in to reply to this topic.
