cancel
Showing results for 
Search instead for 
Did you mean: 
Reply
JWall
New Member

OCR input requirements

Hello,

 

I am attempting to create an AI to recognise custom document information, and I am coming into multiple problems with OCR correctly identifying text. I realise this is a known problem and is being countermeasured with constant updates to the OCR model as per power automate community forum post 'Problem with Model recognising Zero and letter O'

 

I noticed one of the input requirements for OCR is ~8pt font text in order to Read. When analising the document type I am using train the AI model, at standard size, the font is ~8pt 2022-07-29_07-54-05.png

2022-07-29_07-35-59.png

 

 

 

 

 

 

 

 

 

 

 

 

Now, I see where this could be a problem with a document that employs raster imaging, in which the number of pixels in an image is predetermined, and when you zoom in the document appears to have lower resolution. The document type that I am attempting to analyse however appears to be utilising vector imaging, in which shapes are determined by a set of geometrical equations and resolution scales up the more you zoom in

2022-07-29_08-04-312.png2022-07-29_08-06-12.png

 

My question; does the imaging technique for a document have an effect in the ability of OCR to correctly identify text? This issue is not as prevalent with larger style letters within the same document (same imaging technique as smaller style letters)

 

2022-07-29_08-24-50.png

 

 

 

 

 

 

 

 

Appreciate the help!

4 REPLIES 4

Hi @JWall - thanks for the question and the detailed analysis.

 

A few things we can try to see if you see any impact:

  • If you navigate to the homepage of AI Builder (https://aka.ms/tryaibuilder) --> Select Text recognition / Extract all the text in photos and PDF documents (OCR). --> Upload new. Do you get the data extracted as you would expect? The text recognition model has been recently updated with OCR improvements. 

    JoeFMSFT_0-1659194581339.png

     

    JoeFMSFT_1-1659194592015.png

     

  • If you print the original PDF as a new PDF, and use the newly generated PDF on the text recognition model as the step before, do you see better results?

  • If you try to transform one of the pages of the PDF document into an image (for example by taking a screenshot) and try here again with text recognition, do you notice any improvements?

 

JWall
New Member

Hi @JoeF-MSFT

Thanks for the reply!

 

TL;DR - Trying the different methods appeared to have no impact to improving results. Not sure if there are any other methods/variables to test. A suggested feature I could make though would be to allow for manual entries in AI builder, where you still highlight the field in which you want the model to read, but if the model is unable to correctly read the text, then allow for an option to manually edit the read value for the field. Adding incorporation to the MS OCR recognition model to allow for improvement to that as well as improving end user AI models would greatly help the robustness and flexibility of AI builder. 

 

 

Unfortunately I am unable to upload as detailed of a report as last time due to sensitive information, however; I did run through analysis on the situations you suggested. Utilising the 'extract all text in photos and PDF documents (OCR)' default model and uploading my original document, a version of the document that was printed as a new PDF -> SaveAs, and finally a version of the document that was taken as a screenshot and saved as a JPEG. I also tried a version of the document that was taken as a screenshot and saved as a PDF after seeing the results.

From a character count perspective, the results from what the AI reads are as follows:

LEN(.pdforiginal)799
LEN(.jpgss)920
LEN(.pdfprint)799
LEN(.pdfss)500

The original PDF and printing ->saveas PDF yielded the same results. Interestingly; the screenshot -> JPEG had the highest character count, while the screenshot -> PDF had the lowest character count.

 

Now when comparing to the actual data, I am unable to get an exact character count on the original PDF without meticulously counting it myself. What I can tell, is none of the AI read results correctly extracted the data as I would expect. For example the sample document has a total of 18 'A's in a table (similar to that of my previous post). None of the AI read results showed any amount of consistency in 1. Detecting a 'word', 2. Correctly identifying the 'word'. I think at best, surprisingly the jpeg version performed the best at the specific task correctly identifying ~8 'A's, but again; not to adequate result. The original PDF appeared to correctly identify the most amount of characters, which can help explain why the .jpg version identified more characters. A prime example of this would be the .jpg version identifying a column line as an 'l'.

 

Not sure if there are any other methods I could try to help troubleshoot or test for better methods. Other than patiently waiting for improvements to the character recognition AI model. A suggestion I could make though would be to allow for manual entries in AI builder, where you still highlight the field in which you want the model to read, but if the model is unable to correctly read the text, then allow for an option to manually edit the read field value. Adding incorporation to the MS OCR recognition model to allow for improvement to that as well as improving end user AI models would greatly help the robustness and flexibility of AI builder. 

 

Appreciate the help, and let me know if you have any more thoughts. Thanks!

Hi @JWall - I really appreciate the detailed investigations! And thanks for the feedback of allowing to provide feedback on the detected words while tagging the documents. This is something that indeed we don't have today.

 

I'm curious about those 'A's that are not detected. I understand that the documents contain sensitive information. Would it be possible to share just a screenshot of a word where an 'A' is not detected? Or maybe a partial screenshot of that word?

JWall
New Member

Hi @JoeF-MSFT - Sure.

 

For this specific example I have an array of letters in a table. The letters aren't always 'A', nor are they always aligned in a linear layout pattern. The first screenshot helps show an example of a letter not being detected. All 'B's are detected by the OCR software except for the 'B' highlighted in red. The other 'B's that are either not showing up in the table on the right, or misplaced in the table on the right can easily be fixed by moving the column line, and are correctly identified as text by OCR. You may also notice that the array has '.' in some of the fields. Sometimes these are detected, and sometimes they are not and to which is varying degrees of success. I am not so concerned with this as '.' can also be treated as a blank in my use case, but you may find this interesting for your use.3.png

 

4.png

 

 

 

 

This is a slightly different example where the OCR is detecting a column line as text ('|'). Sometimes OCR will detect column lines as 'l'.

 

 

 

 

 

Again, I want to highlight I presumed this problem to possibly be due to the OCR not being accurate with font sizes <=~8pt, but was curious as to if the image processing type would have an effect on that (vector vs raster) (these PDFs used were vector). 

On an alternate note, I am going to attempt to work around the need to use OCR. These documents I was trying to use with OCR are all standard tables internal to our company. The thing is that they are 1. versions of pivot tables that make it easier for a human to read, 2. in PDF format and thus not as easy for a computer to read (the problem I was trying to solve with OCR). I am working with some people in my company to gain some additional information, but I would go to think there is some more raw data that are driving these PDF documents (strings in an array, tabular format or something like that), something which a computer may have a bit of an easier time reading. If this information exists and I can get access to it, then I should be able to skip over the process of running through OCR & creating an AI to recognise custom document information. 

 

Anyways, I appreciate the help, and hopefully I was able to help with your inquiry.

Helpful resources

Announcements

Community will be READ ONLY July 16th, 5p PDT -July 22nd

Dear Community Members,   We'd like to let you know of an upcoming change to the community platform: starting July 16th, the platform will transition to a READ ONLY mode until July 22nd.   During this period, members will not be able to Kudo, Comment, or Reply to any posts.   On July 22nd, please be on the lookout for a message sent to the email address registered on your community profile. This email is crucial as it will contain your unique code and link to register for the new platform encompassing all of the communities.   What to Expect in the New Community: A more unified experience where all products, including Power Apps, Power Automate, Copilot Studio, and Power Pages, will be accessible from one community.Community Blogs that you can syndicate and link to for automatic updates. We appreciate your understanding and cooperation during this transition. Stay tuned for the exciting new features and a seamless community experience ahead!

Check Out | 2024 Release Wave 2 Plans for Microsoft Dynamics 365 and Microsoft Power Platform

On July 16, 2024, we published the 2024 release wave 2 plans for Microsoft Dynamics 365 and Microsoft Power Platform. These plans are a compilation of the new capabilities planned to be released between October 2024 to March 2025. This release introduces a wealth of new features designed to enhance customer understanding and improve overall user experience, showcasing our dedication to driving digital transformation for our customers and partners.    The upcoming wave is centered around utilizing advanced AI and Microsoft Copilot technologies to enhance user productivity and streamline operations across diverse business applications. These enhancements include intelligent automation, AI-powered insights, and immersive user experiences that are designed to break down barriers between data, insights, and individuals. Watch a summary of the release highlights.    Discover the latest features that empower organizations to operate more efficiently and adaptively. From AI-driven sales insights and customer service enhancements to predictive analytics in supply chain management and autonomous financial processes, the new capabilities enable businesses to proactively address challenges and capitalize on opportunities.    

Summer of Solutions | Week 3 Results | Win free tickets to the Power Platform Conference

We are excited to announce the Summer of Solutions Challenge!    This challenge is kicking off on Monday, June 17th and will run for (4) weeks.  The challenge is open to all Power Platform (Power Apps, Power Automate, Copilot Studio & Power Pages) community members. We invite you to participate in a quest to provide solutions to as many questions as you can. Answers can be provided in all the communities.    Entry Period: This Challenge will consist of four weekly Entry Periods as follows (each an “Entry Period”)   - 12:00 a.m. PT on June 17, 2024 – 11:59 p.m. PT on June 23, 2024 - 12:00 a.m. PT on June 24, 2024 – 11:59 p.m. PT on June 30, 2024 - 12:00 a.m. PT on July 1, 2024 – 11:59 p.m. PT on July 7, 2024 - 12:00 a.m. PT on July 8, 2024 – 11:59 p.m. PT on July 14, 2024   Entries will be eligible for the Entry Period in which they are received and will not carryover to subsequent weekly entry periods.  You must enter into each weekly Entry Period separately.   How to Enter: We invite you to participate in a quest to provide "Accepted Solutions" to as many questions as you can. Answers can be provided in all the communities. Users must provide a solution which can be an “Accepted Solution” in the Forums in all of the communities and there are no limits to the number of “Accepted Solutions” that a member can provide for entries in this challenge, but each entry must be substantially unique and different.    Winner Selection and Prizes: At the end of each week, we will list the top ten (10) Community users which will consist of: 5 Community Members & 5 Super Users and they will advance to the final drawing. We will post each week in the News & Announcements the top 10 Solution providers.  At the end of the challenge, we will add all of the top 10 weekly names and enter them into a random drawing.  Then we will randomly select ten (10) winners (5 Community Members & 5 Super Users) from among all eligible entrants received across all weekly Entry Periods to receive the prize listed below. If a winner declines, we will draw again at random for the next winner.  A user will only be able to win once overall. If they are drawn multiple times, another user will be drawn at random.  Individuals will be contacted before the announcement with the opportunity to claim or deny the prize.  Once all of the winners have been notified, we will post in the News & Announcements of each community with the list of winners.   Each winner will receive one (1) Pass to the Power Platform Conference in Las Vegas, Sep. 18-20, 2024 ($1800 value). NOTE: Prize is for conference attendance only and any other costs such as airfare, lodging, transportation, and food are the sole responsibility of the winner. Tickets are not transferable to any other party or to next year’s event.   ** PLEASE SEE THE ATTACHED RULES for this CHALLENGE**   Week 1 Results: Congratulations to the Week 1 qualifiers, you are being entered in the random drawing that will take place at the end of the challenge.   Community MembersNumber SolutionsSuper UsersNumber Solutions Deenuji 9 @NathanAlvares24  17 @Anil_g  7 @ManishSolanki  13 @eetuRobo  5 @David_MA  10 @VishnuReddy1997  5 @SpongYe  9JhonatanOB19932 (tie) @Nived_Nambiar  8 @maltie  2 (tie)   @PA-Noob  2 (tie)   @LukeMcG  2 (tie)   @tgut03  2 (tie)       Week 2 Results: Congratulations to the Week 2 qualifiers, you are being entered in the random drawing that will take place at the end of the challenge. Week 2: Community MembersSolutionsSuper UsersSolutionsPower Automate  @Deenuji  12@ManishSolanki 19 @Anil_g  10 @NathanAlvares24  17 @VishnuReddy1997  6 @Expiscornovus  10 @Tjan  5 @Nived_Nambiar  10 @eetuRobo  3 @SudeepGhatakNZ 8     Week 3 Results: Congratulations to the Week 3 qualifiers, you are being entered in the random drawing that will take place at the end of the challenge. Week 3:Community MembersSolutionsSuper UsersSolutionsPower Automate Deenuji32ManishSolanki55VishnuReddy199724NathanAlvares2444Anil_g22SudeepGhatakNZ40eetuRobo18Nived_Nambiar28Tjan8David_MA22  

Updates to Transitions in the Power Platform Communities

We're embarking on a journey to enhance your experience by transitioning to a new community platform. Our team has been diligently working to create a fresh community site, leveraging the very Dynamics 365 and Power Platform tools our community advocates for.  We started this journey with transitioning Copilot Studio forums and blogs in June. The move marks the beginning of a new chapter, and we're eager for you to be a part of it. The rest of the Power Platform product sites will be moving over this summer.   Stay tuned for more updates as we get closer to the launch. We can't wait to welcome you to our new community space, designed with you in mind. Let's connect, learn, and grow together.   Here's to new beginnings and endless possibilities!   If you have any questions, observations or concerns throughout this process please go to https://aka.ms/PPCommSupport.   To stay up to date on the latest details of this migration and other important Community updates subscribe to our News and Announcements forums: Copilot Studio, Power Apps, Power Automate, Power Pages

Top Solution Authors
Users online (4,774)