Document Understanding - Checkboxes and Signatures

Document Understanding - Checkboxes and Signatures (uipath.com)
The documentation above suggests that we can create a single field for checkboxes so that the data extracted is the option or options selected.
I can’t find any good examples for extracting the text indicated by a checkbox, like this example below from the “Build a Document Understanding Automation in Studio” course in the Specialized AI Professional learning plan.
image
In this example, I would want the “Card Type” field to extract the text “Domestic Card.” If someone has the other field checked we would want “International Card” extracted.

This example correctly identifies if an option is selected or not, but it separates out all the choices, so they aren’t connected in any way.

Is there a workflow example out there of creating a single field that correctly picks the text of the option that is next to the checkbox, like what is described below?

I can figure this out in time, but it would be easiest to have a tangible example of this being done. Thanks.

@Joe_Matuch

I believe you need to follow this…where you would try to get documents equally checked for two fields and label them…with only name beside which the check is present…and represent the fields with a multi value field and link both to same

Cheers

That sounds great and is what I am trying to do. Sometimes the documentation includes a sample workflow or steps for accomplishing a task. In this case, I don’t see one, so unless anyone has an example they’d like to share, I will try to figure this out myself.

@Joe_Matuch

Sorry when I said this I was saying about the steps mentioned in the link shared by you for the first example

We did try a similar one…but it is on our company laptop…will check if I can get soem references

Cheers

No problem. Thanks for your help. I think for now using a Form Extractor activity, I will either

  1. Identify each checkbox separately as a Boolean and use an If/Else to assign what text should be assigned to a variable, or
  2. Extract the entire field, like “Domestic ☒ International ☐” and do string manipulation to determine which option is selected.

I would lean toward the first option since it’s possible for someone to submit a form with symbols other than X in the checkbox.

Of course, there may be methods to do this more accurately using an ML package and skill, but I haven’t tested that out yet.

1 Like

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

Please note that with IntelligentOCR you can manipulate Everything! Meaning, you can do data extraction on a taxonomy that contains two fields (like Domestic / International), but you can then, based on the results, play around and generate a value for a “combined” field (say Type of Card - set field with the two options), based on whatever findings you actually get from the individual extractions… I hope this makes sense :slight_smile: