'Read PDF With OCR' is not reading line by line from PDF

Help Studio
1

Hi,

I am extracting text from image in pdf with ‘Read PDF with ocr’ activity with ‘UiPath Extended Languages OCR’ engine.

I am getting below result
456
123
789

But, I am expecting read line by line in pdf and looking to get output as
123
456
789

Appreciate any help.

image
image212×133 397 Bytes

2

Method 1: Configuring OCR Engine Settings

1. Scale setting has been adjusted in the properties panel.
• The scale value usually changes to 1.0, but some may use higher values ​​such as 1.5 or 2.0, which can help it read more accurately.
2. Inversion option option.
• If the background color of the text in the PDF is dark, you can enable the Invert option so that the OCR engine detects the text better.

Method 2: Replace OCR Engine

• Apart from Extended Languages ​​OCR, try Tesseract OCR, Microsoft OCR or Google OCR engines.
• For each OCR engine to run differently, it is possible to sequence some copies more accurately with another engine.
3

Instead of using Read PDF with OCR use Digitize activity from IntelligentOCR package.

It provides a string output much more similar to the document layout

1 Like
4

Tried Digitize as well, didn’t work. Thanks.

5

Tried different free OCRs. ‘UiPath Extended Languages OCR’ is most suitable for my requirement, but it is not giving sequential line by line output. Thanks

6

@chandrasekhar.jella

If you want to use read PDF Text, It can be help to you.It will not change format

The image shows a workflow for reading a PDF text in a software automation environment, highlighting the option to preserve formatting with the "PreserveFormatting" parameter set to "True." (Captioned by AI)
The image shows a workflow for reading a PDF text in a software automation environment, highlighting the option to preserve formatting with the "PreserveFormatting" parameter set to "True." (Captioned by AI)1114×423 17.9 KB

7

Hi @chandrasekhar.jella

If none of them work and you can’t use the paid feature, you can try using the traditional method:
1-) Open the PDF and grab all the text with Ctrl+A and set it to the clipboard (Ctrl+C)
2-) Use the Get From Clipboard activity

You can get the text this way and check it. I can’t say it’s a definitive solution, but I’ve had this problem before and it was solved with this method :slight_smile: