Computer vision Find CV Text

Help
1

Hello,
I am using computer vision type into to write into a text box. The text box is labeled as “Description” but computer vision activity identifies it as “Dæcrjpbon:”. I ran the workflow couple of times and it seems to be working fine but should I change the name to “Description” to find cv text? Please see attached snapshots.
2019-07-09_9-36-24

2019-07-09_9-36-55

Thanks in advance for your reply.
Rishi

2

Hi @RISHI4897

I had similar Scenario while Typing into a Password.
It read the spelling wrong and works perfectly fine.

I tried correcting the Spelling manually but then it did not work.
So I am going with the wrong spelling unless it gives me the desired value :sweat_smile:

It is basically OCR, if the Image resolution changes next time then it can give a different spelling.

3

Hi @RISHI4897

The OCR engine of the computer vision component use a specific scale to read the data out of the screen, so if you try change the value manually it won’t match it since the specific scale you’re working with is detecting Description as Dæcrjpbon.

My suggestion is to let it as it is since it’s always going to work, or you can change the scale of the Microsoft OCR engine so it can get corrected by itself, of even change the engine to Tesseract OCR, But this will affect all the things you’ve already done.

I mean you don’t need to bother really, you know… “If it’s working than don’t touch it”.

Regards,
Reda

4

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.