web

You’re offline. This is a read only version of the page.

Power Platform CommunityForums Power Automate - General Discussion PDF "Forms Processing"...

Power Automate - General Discussion

Unanswered

PDF "Forms Processing" ocr missing text

Like (0) Share

Report

Report

Posted on 6 May 2021 12:51:49 by GusG

Hi,

Very new to this functionality.

I just finished developing a model (with 7 samples used to train) of a Sales Order (SO)received in a PDF format.

1. During the training the OCR missed several simple numbers/characters in table line items and was unable to select them manually

2. During testing (see below) it seems that the OCR was also unable to recognize all numbers and text in table.

Is there any way for me to debug why this is happening? I also, performed a search on the PDF and it appears that the form recognizes the simple text/numbers missed by the Azure ORC.

Any help would be appreciated.

Gus

I have the same question (0)

All responses Img

All responses (1)

Answers Img

Answers (0)

JoeF-MSFT on 09 May 2021 at 13:36:57

Like (0)

Report

Re: PDF "Forms Processing" ocr missing text

Hi @GusG,

Thanks for sharing about this.

Today the OCR engine we use in AI Builder form processign can at times have a hard time recognizing single characters in documents. This seems to be the issue in this case.

We will be switching form processing to a newest, more robust OCR engine in July that will be able to handle this.

Was this reply helpful? Yes No

Join the conversation

Helpful resources

News and Announcements

Welcome to the Power Platform Communities

Quick Links

Leaderboard > Power Automate

#1

Michael E. Gernaey 788 Super User 2025 Season 2

#2

Tomac 452 Moderator

#3

developerAJ 302

Last 30 days Overall leaderboard

Featured topics

Survey: Power Automate mobile app

How to Ask for Help with Your Power Automate Workflow

Product updates

Microsoft Power Platform Community release plans