Unanswered

Extraction of Repeating Fields in a pdf using AI Builder-Form Processing Model

(0) Share

Report

Posted on by Prajakta05

I have pdfs in which a single field is present more than once, since field value each time is different, I'm trying to the tag all the field values. Is it possible to tag more than one field value for a field in AI Builder ?
If No, what would be the alternate option ?

Categories:

AI Builder

I have the same question (0)

All responses (7)

Answers (0)

antoinec on at

Like (0)

Report

Hi Prakak, if what you are trying to extract is a list of different items with the same structure, you probably want to look into tables: https://docs.microsoft.com/en-us/ai-builder/create-form-processing-model#tag-tables. Any item with repeating structure can be extracted using a table even if it doesn't not look like a table. Do let us know if that works,

Antoine

Was this reply helpful? Yes No
Prajakta05 86 on at

Like (0)

Report

Hello,
I tried capturing repeating fields in table format using single page table. I wanted to extract table from 2 pages so, I added 2 tables named as Location schedule 1 & Location schedule 2. After training the model for 17 pdfs, it predicted both the tables on the same location.
Help me with this.

Was this reply helpful? Yes No
antoinec on at

Like (0)

Report

Which of the values are you looking to extract? Could you highlight them on a screenshot to help understand what you're looking to achieve? Thanks!

Was this reply helpful? Yes No
Prajakta05 86 on at

Like (0)

Report

Hello,
I'm trying to extract set of fields called 'Location Schedule' from page 2 & 3. I've tried extracting these sets as table using advanced tagging, but it predicts incorrectly.
I've attached screenshot for your ready reference.
Thanks in advance !

Table 1 from page 2

Table 2 from page 3

Was this reply helpful? Yes No
JoeF-MSFT on at

Like (0)

Report

Hi @Prajakta05 - thanks for the additional info!

What I would recommend, is to use the following flow template that will process the document page by page: https://docs.microsoft.com/en-us/ai-builder/form-processing-multipage#use-a-cloud-flow-to-process-all-pages-in-the-document You will see in this page the link to the template as well as the instructions on how to use it.

By processing page by page, the model will be able to correctly extract the table in each page and at the end of the flow execution you will have captured all tables.

Keep us posted if this approach works for you or not.

Was this reply helpful? Yes No
somboNR 2 on at

Like (0)

Report

Edit:
Didn't realize we can specify pages for AI Builder to scan.
Page Range

We can get how many pages in the PDF from this link.
PDF Pages in Power Automate

We can loop the document processor according to the number of pages later

Hope this helps others looking for this.

Hi @JoeF-MSFT ,

How do you propose we do this page by page? I couldn't find a free way to split the pdf in the flow.

Thanks

Was this reply helpful? Yes No
takolota1 4,974 Moderator on at

Like (0)

Report

You could also try this template using AI Builder to extract any PDF data to a JSON object:

https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Extract-Data-From-PDFs-and-Images-With-GPT/td-p/2201345

Was this reply helpful? Yes No