OCR extract data fail

[And9612]

Hello, 

I have a problem whit extract data from an ocr image, i use an application to save image whit the ocr data, i save the print screen whit data, by using that applicatin, but when i need to extract the ocr(captcha) and to paste in specify box on a website it don t work, do not fill the box, not populate that box, i try to understand this situation i thought first it s a problem with populate i disable troggle button with physical keystrokes, but i don t know why do not fill that box or maybe not extract correctly ocr...

Thank you!

this is the flow

[VishnuReddy1997]

Hi @And9612 ,

are you getting any output from extract text with ocr action.

Can you please share the image of the input that your provide to extrat with ocr.

Regards,

Vishnu Reddy

[eetuRobo]

Does the OcrText variable have some text in it when your flow runs Populate text field on web page -action? Also does it have just one line of text of multiple lines? Try to use Trim -action on it first so that you would trim empty lines from the begining and the end.

Also maybe the text field on that web page wants singe line text and your OcrText has multiple lines of text and that's why it doesn't populate.

[And9612]

this is the image after collor corection and before extracte ocr, after that it not populate in the box

[eetuRobo]

So does that not return any value from the OCR action? If OcrText -variable is empty then of course it wont populate anything in the web pages text field since there is no value to be populated. If OcrText variable is empty it wasn't able to read the image.

Also captchas are deliberately made to be difficult to be read by a computer so you might not be able to automate this.

[VishnuReddy1997]

Hi @And9612 ,

Your not try to understand what we are asking, we want the image of Extract text with OCR action after running the flow as shown in below images.

We wanted to see what that action is extacting from the image.

Regards,

Vishnu Reddy

[And9612]

Maybe the site where the OCR result must be uploaded has had an update lately and because of this it doesn't work? Previously it worked 

[And9612]

also it return just one or 2 results, but after that it do not return value in the box

[VishnuReddy1997]

Hi @And9612 ,

Can send the image ,so we will try with Extract text with OCR action and come back to you.

Regards,

Vishnu Reddy

[eetuRobo]

Yeah might have had improvement update so the captcha is not so easily readable with OCR. 
What do you mean it returns just one or 2 results? That it returns only 1 or 2 first numbers/letters of the captcha?

[Deenuji]

@And9612 

If you find the default PAD OCR insufficient for extracting text from your image, you might consider using Microsoft Cognitive Services' Computer Vision OCR action in PAD, or you could utilize AI Hub for improved results.

Thanks,
Deenuji Loganathan 👩‍💻
Automation Evangelist 🤖
Follow me on LinkedIn 👥

-------------------------------------------------------------------------------------------------------------
If I've helped solve your query, kindly mark my response as the solution ✔ and give it a thumbs up!👍 Your feedback supports future seekers 🚀