[Athen] Handwriting to OCR text

[Joshua Hori via athen-list]

I have not tried them. The pricing seems high for usage, but I would be interested to see if the API is a cheaper solution.

Best,

Joshua

From: Kamran Rasul <krasul1 at jhu.edu>
Date: Tuesday, April 15, 2025 at 12:10 PM
To: Joshua Hori <jhori at ucdavis.edu>, 'Access Technology Higher Education Network' <athen-list at u.washington.edu>, 'enews at toptechtidbits.com' <enews at toptechtidbits.com>, 'shelleyhaven techpotential.net' <shelleyhaven at techpotential.net>, 'Karen McCall' <K4mccall at outlook.com>
Subject: RE: [Athen] Handwriting to OCR text
Have you tried giving this a try? Although I cannot speak for math content.

https://www.handwritingocr.com/

[Johns Hopkins University logo]
Kamran Rasul, MEd.
Assistive Technology/Alternate Format Specialist (SDS)
Phone: 410-516-1167
E-mail: krasul1 at jhu.edu<mailto:krasul1 at jhu.edu>
Garland Hall, 1st Floor, Office 135-G
3400 N. Charles Street
Baltimore, MD 21218
Schedule a meeting with Kamran<https://outlook.office365.com/owa/calendar/ATC2@live.johnshopkins.edu/bookings/>

From: athen-list <athen-list-bounces at mailman12.u.washington.edu> On Behalf Of Joshua Hori via athen-list
Sent: Saturday, April 12, 2025 7:49 PM
To: enews at toptechtidbits.com; 'Access Technology Higher Education Network' <athen-list at u.washington.edu>; 'shelleyhaven techpotential.net' <shelleyhaven at techpotential.net>; 'Karen McCall' <K4mccall at outlook.com>
Subject: Re: [Athen] Handwriting to OCR text


      External Email - Use Caution




We have https://aggievideo.ucdavis.edu<https://aggievideo.ucdavis.edu/> which has many public STEM courses you can test with me. 😃
Following up on the math tests:

Math equations proved to be interesting, but I’m testing with multivariable calculus…Which has math characters most will never see. Here’s a whiteboard pulled from the course:

[cid:image004.png at 01DBAE18.7884AA50]


  *   Mathpix gets most of it but needs a little readjusting. Just a little.
  *   Gemini 2.5 Pro: gets quite a bit but messes up on the “notion & def” line, all of it.

     *   Uses \bar(x) instead of \vec(x)
     *   Uses \geq d instead of \geqslant d
     *   Also has some mathml code showing instead of proper math output. Pops up in different places if regenerating the response.

  *   GPT o3-mini-high: gets almost everything, but leaves out the “\forall x,y,z”, lumping that last line with the previous “for all”. It also:

     *   Uses \bar(x) instead of \vec(x)
     *   Uses \geq d instead of \geqslant

  *   Claude Sonnet 3.7: Sections the equations, but gets is sectioned correctly.

     *   Uses \bar(x) instead of \vec(x)
     *   Uses \geq instead of \geqslant

  *   Perlexity: Got everything right, with a few minor edits here and there. Less than Mathpix edits. *Shocked pikachu*.

     *   Uses \geq instead of \geqslant

  *   Copilot…tries. It tends to rewrite the equation.

     *   \bar? \vec? We don’t do that here…
     *   So many missing symbols, so many missing characters…

\forall = upside down A
\geq = greater than or equal to
\geqslant = greater than or equal to, slant
\bar(x) = x with bar over it
\vec(x) = x with arrow over it

Best,

Joshua

From: Top Tech Tidbits <enews at toptechtidbits.com<mailto:enews at toptechtidbits.com>>
Date: Saturday, April 12, 2025 at 1:27 PM
To: Joshua Hori <jhori at ucdavis.edu<mailto:jhori at ucdavis.edu>>, 'Access Technology Higher Education Network' <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>, 'shelleyhaven techpotential.net' <shelleyhaven at techpotential.net<mailto:shelleyhaven at techpotential.net>>, 'Karen McCall' <K4mccall at outlook.com<mailto:K4mccall at outlook.com>>
Subject: RE: [Athen] Handwriting to OCR text
Thanks so much for the mention Joshua, but I will defer to your review which is much more thorough. We too tested this across all of the large paid models and ChatGPT 4o and Gemini 2.5 Pro consistently performed the best.

Aaron Di Blasi, PMP<https://www.pmi.org/>

[Aaron Di Blasi, Sr. Project Management Professional, Mind Vault Solutions, Ltd., Publisher, Top Tech Tidbits, Publisher, Access Information News stands smiling, arms crossed, in a suit and tie.]<https://www.linkedin.com/in/aarondiblasi/>

“The greatest barrier to accessibility is indifference.” 💡

PR Director (2024-Present)
AT-Newswire
Access Technology's Digital Newswire
https://at-newswire.com<https://at-newswire.com/> 🌐

Publisher (2024-Present)
AI-Weekly
The Week's News in Artificial Intelligence
https://ai-weekly.ai<https://ai-weekly.ai/> 🌐

Publisher (2022-Present)
Access Information News
The Week's News in Access Information
https://accessinformationnews.com<https://accessinformationnews.com/> 🌐

Publisher (2020-Present)
Top Tech Tidbits
The Week's News in Access Technology
https://toptechtidbits.com<https://toptechtidbits.com/> 🌐

Sr. Project Management Professional (2006-Present)
Mind Vault Solutions, Ltd.
https://mvsltd.com<https://mvsltd.com/> 🌐

Certified:
Digital Marketing Associate, Meta Certified<https://mvsltd.com/news/aaron-di-blasi-pmp-mind-vault-solutions-ltd-awarded-digital-marketing-associate-certification-by-meta/> (2022 - Present)
Social Marketing Professional, Hootsuite Certified<https://mvsltd.com/news/aaron-di-blasi-pmp-mind-vault-solutions-ltd-awarded-social-media-marketing-certification-by-hootsuite-world-leader-in-social-media-marketing-solutions/> (2020 - Present)
Email Marketing Professional, Constant Contact Certified<https://mvsltd.com/news/aaron-di-blasi-pmp-mind-vault-solutions-ltd-named-a-constant-contact-certified-solution-provider/> (2019 - Present)

Specializing in:
Digital Strategy and Content Marketing
Social Media Advertising
Online Fundraising
ADA<https://www.ada.gov/>, WCAG<https://www.w3.org/WAI/standards-guidelines/wcag/> and Section 508 Compliance<https://www.justice.gov/crt/section-508-home-page-1>

Website: https://mvsltd.com<https://mvsltd.com/> 🌐
Email: ad at mvsltd.com<mailto:ad at mvsltd.com> 📧
Toll Free: +1 (855) 578-6660<tel:+18555786660>📱️

Schedule A Meeting: https://calendly.com/aarondiblasi
News: https://mvsltd.com/news
Services: https://mvsltd.com/services
Testimonials: https://mvsltd.com/testimonials
Facebook: https://mvsltd.com/facebook
X (Formerly Twitter): https://mvsltd.com/x
LinkedIn: https://mvsltd.com/linkedin
Instagram: https://mvsltd.com/instagram
YouTube: https://mvsltd.com/youtube
Google: https://mvsltd.com/google

CONFIDENTIALITY NOTICE: This e-mail and attachments, if any, may contain confidential information, which is privileged and protected from disclosure by Federal and State confidentiality laws, rules, and regulations. This e-mail and attachments, if any, are intended for the designated addressee only. If you are not the designated addressee, you are hereby notified that any disclosure, copying, or distribution of this e-mail and its attachments, if any, may be unlawful and may subject you to legal consequences. If you have received this e-mail and attachments in error, please delete the e-mail and its attachments from your computer.

From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> On Behalf Of Joshua Hori via athen-list
Sent: Friday, April 11, 2025 5:38 PM
To: Access Technology Higher Education Network <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>; shelleyhaven techpotential.net <shelleyhaven at techpotential.net<mailto:shelleyhaven at techpotential.net>>; Karen McCall <K4mccall at outlook.com<mailto:K4mccall at outlook.com>>
Subject: Re: [Athen] Handwriting to OCR text

And now I’ve tested with AI:

Copilot: Gives the handwritten text, nothing more.
ChatGPT: Gives the handwritten text, then explains it. The handwritten text is displayed through a code window that can be quickly copied separate from the explanation.
Claude: Gives the handwritten text but is messy and needs additional formatting. It sticks multiple bullet points into one line.
Perplexity: Gives the handwritten text and explanation but gets some of the latin font wrong.
Gemini: Gives the handwritten text and explanation, one of the better solutions.

I’ll have to test with some math equations next, then some chemistry lectures.

Oh look! An email came in and I see Aaron from Top Tech is stating the same thing. 😃

Best,

Joshua

From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> on behalf of Joshua Hori via athen-list <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Date: Friday, April 11, 2025 at 1:57 PM
To: shelleyhaven techpotential.net <shelleyhaven at techpotential.net<mailto:shelleyhaven at techpotential.net>>, Karen McCall <K4mccall at outlook.com<mailto:K4mccall at outlook.com>>, Access Technology Higher ATHEN <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Subject: Re: [Athen] Handwriting to OCR text
Thanks for providing this explanation, Shelley! I appreciate the detailed comparison between handwriting OCR and dynamic handwriting recognition.

I just tested it out in OneNote, you can’t lasso handwriting in images. If an image is taken of a whiteboard with handwriting, it will not work. If you rewrite the handwritten notes, you can lasso that.

Best,
Joshua

From: shelleyhaven techpotential.net <shelleyhaven at techpotential.net<mailto:shelleyhaven at techpotential.net>>
Date: Friday, April 11, 2025 at 1:01 PM
To: Karen McCall <K4mccall at outlook.com<mailto:K4mccall at outlook.com>>, Access Technology Higher ATHEN <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Cc: Joshua Hori <jhori at ucdavis.edu<mailto:jhori at ucdavis.edu>>, Karlen Communications <info at karlencommunications.com<mailto:info at karlencommunications.com>>, Travis Roth <travis at travisroth.com<mailto:travis at travisroth.com>>
Subject: Re: [Athen] Handwriting to OCR text
I've been following this handwriting OCR thread with great interest.

While not a direct answer to Lorraine's original question (converting a PDF of handwritten notes to OCR'd text), here's one other thing to consider relevant to accommodations: there's a difference between handwriting OCR and dynamic handwriting recognition.

Handwriting OCR examines the static, after-the-fact image of handwritten text or math and attempts to recognize what the user wrote by analyzing the shapes and making a "best guess" by comparing these shapes with known handwriting samples.  The process is similar to the standard OCR we know and (mostly) love.

Dynamic handwriting recognition, however, analyzes handwriting in real time (as it's being written) and as such has access to additional information with which to make decisions: the individual strokes used to create the handwriting, the sequence of the strokes, the direction of the writing utensil, the velocity of the writing utensil, pen-up and pen-down movements, and possibly the pressure of the strokes.  This produces far more accurate handwriting recognition because it's analyzing "how the image was made", not just "what the image looks like".

Dynamic HWR obviously requires a system to capture the handwriting in real time, like a digitizing tablet or touchscreen.  OneNote on a touchscreen, Livescribe smartpens, Scribble on iPad with Apple Pencil, and many other apps & devices use dynamic HWR.

So in terms of accurately recognizing and converting a student's handwriting to text or math, it is advantageous to employ an accommodation that allows the student to capture their handwriting in real time using a dynamic HWR system.

- Shelley

_____________________________
Shelley Haven  ATP, RET, BSME
Assistive Technology Consultant
www.TechPotential.net<http://www.techpotential.net/>



On Apr 11, 2025, at 5:53 AM, Karen McCall via athen-list <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>> wrote:

Yes.

I had to charge my Surface Pen so used the touch screen to “write” a sentence. I then used the Lasso tool to select the text I’d written.

I tapped on one of the handles of the lasso “square” and got the toolbar. There is an arrow in the toolbar which I tapped. The last item in the context menu is Ink to Text.

It worked perfectly. Even with my handwriting.

Once my Surface pen is charged, I’ll try it with that.

The way the handwriting tool used to work in OneNote is based on algorithms of thousands of handwriting sample, so it is sort of like word prediction based on the database of handwriting samples. Of course, like voice recognition, it also helps you write better/more legibly as you see the common “oops, I got that wrong” results if any. So, it tends to train you instead of the other way around.

Cheers, Karen

From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> On Behalf Of Joshua Hori via athen-list
Sent: Thursday, April 10, 2025 8:13 PM
To: Karlen Communications <info at karlencommunications.com<mailto:info at karlencommunications.com>>; 'Access Technology Higher Education Network' <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>; 'Travis Roth' <travis at travisroth.com<mailto:travis at travisroth.com>>
Subject: Re: [Athen] Handwriting to OCR text

I’ve tried it and couldn’t get it to work, so I abandoned the idea for a while. I then realized later that you need to use the lasso to convert the handwriting to text.

So…Did the lasso convert imaged handwriting to text?

Joshua


From: Karlen Communications <info at karlencommunications.com<mailto:info at karlencommunications.com>>
Date: Wednesday, April 9, 2025 at 5:40 AM
To: Joshua Hori <jhori at ucdavis.edu<mailto:jhori at ucdavis.edu>>, 'Access Technology Higher Education Network' <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>, 'Travis Roth' <travis at travisroth.com<mailto:travis at travisroth.com>>
Subject: RE: [Athen] Handwriting to OCR text
Have you checked OneNote’s ability to convert handwriting to text? Not sure it would work on a scanned graphic of both handwriting and print but might be worth a try.

I used to use it all the time for note taking. Even with my horrid handwriting, it converted accurately.

…and you’ve just given me a task to test this morning!

Cheers, Karen

From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> On Behalf Of Joshua Hori via athen-list
Sent: Tuesday, April 8, 2025 8:59 PM
To: Travis Roth <travis at travisroth.com<mailto:travis at travisroth.com>>; Access Technology Higher Education Network <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Subject: Re: [Athen] Handwriting to OCR text

The camera fails and throws errors now but if you hover over it, it will still try to OCR and read it aloud. We used to use SeeingAI back in 2018 to convert handwritten notes into digital text. It worked with good accuracy back then.

This is the Image Braille GPT (free): https://chatgpt.com/g/g-fXlS5yyJA-image-braille

It’s a little challenging navigating OpenAI with a screenreader since they don’t have ARIA live tags. I find Gemini more accessible to screenreaders. NotebookLM is accessible…with exception to navigating their chat responses.

Best,

Joshua Hori
Accessible Technology Coordinator
Information Educational Technology
Academic Technology Services
50 Hutchison Dr.
Davis, CA 95616
530-752-2439
Schedule a meeting via Calendly<https://calendly.com/d/ytt-hsj-vbn>



From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> on behalf of Travis Roth via athen-list <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Date: Tuesday, April 8, 2025 at 5:34 PM
To: 'Access Technology Higher Education Network' <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Subject: Re: [Athen] Handwriting to OCR text
Hi Joshua,
On the Read tab SeeingAI does have a take Picture button which will still show the document view. But I’ve not tried handwriting with it, so if you are saying even the take picture mode is bad at handwriting that is unfortunate.
Can you share how I may find/try the Image Braille agent?


From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> On Behalf Of Joshua Hori via athen-list
Sent: Tuesday, April 8, 2025 7:18 PM
To: foreigntype at gmail.com<mailto:foreigntype at gmail.com>; Access Technology Higher Education Network <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>;lorraine at rcn.com<mailto:lorraine at rcn.com>
Cc: Access Technology Higher Education Network <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Subject: Re: [Athen] Handwriting to OCR text

SeeingAI used to convert handwriting to text, but now it just attempts to read it aloud…and it’s not as good as it used to be. That’s a shame.

You can upload an image of the handwritten text to AI and have it converted to digital text.

Gemini: Best AI for image descriptions and handwritten text.
ChatGPT: Creates great images but can go wildly inaccurate unless using the “Image Braille” GPT. That’s some high-quality system prompting powering that customized GPT.

I have video examples of using Gemini and OpenAI on my mobile device to describe lecture content in a LinkedIn article: https://www.linkedin.com/pulse/ai-musings-bridging-tactile-chemistry-lecture-audio-joshua-hori-kyglc/<https://www.linkedin.com/pulse/ai-musings-bridging-tactile-chemistry-lecture-audio-joshua-hori-kyglc/?trackingId=sKmgtHt1SPWYrhAnngJuVA%3D%3D>

Sorry, no video or audio editing tools available. I stream/record on a budget.

Best,

Joshua Hori
Accessible Technology Coordinator
Information Educational Technology
Academic Technology Services
50 Hutchison Dr.
Davis, CA 95616
530-752-2439
Schedule a meeting via Calendly<https://calendly.com/d/ytt-hsj-vbn>




From: athen-list <athen-list-bounces at mailman12.u.washington.edu<mailto:athen-list-bounces at mailman12.u.washington.edu>> on behalf offoreigntype at gmail.com<mailto:foreigntype at gmail.com> via athen-list <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Date: Tuesday, April 8, 2025 at 8:37 AM
To: lorraine at rcn.com<mailto:lorraine at rcn.com> <lorraine at rcn.com<mailto:lorraine at rcn.com>>
Cc: Access Technology Higher Education Network <athen-list at u.washington.edu<mailto:athen-list at u.washington.edu>>
Subject: Re: [Athen] Handwriting to OCR text

I am unaware of any OCR software that can do both handwriting and text.

Susan's advice is the best: You'll have to do this in several steps. Use a good text based OCR software (Abbyy FineReader or OmnipagePro) for the text and insert (type up the handwriting manually) separately after removing it (by cleaning up) from the original.

No easy button for this one!

Wink

On Tue, Apr 8, 2025, 7:20 AM <lorraine at rcn.com<mailto:lorraine at rcn.com>> wrote:
Wink,

It is handwriting on a PDF.  It is the other software that I am looking for recommendations on.

Thanks

Lorraine
_______________________________________________
athen-list mailing list
athen-list at mailman12.u.washington.edu<mailto:athen-list at mailman12.u.washington.edu>
http://mailman12.u.washington.edu/mailman/listinfo/athen-list

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20250415/9fc46e85/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.png
Type: image/png
Size: 8704 bytes
Desc: image003.png
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20250415/9fc46e85/attachment-0003.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image004.png
Type: image/png
Size: 441256 bytes
Desc: image004.png
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20250415/9fc46e85/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image005.png
Type: image/png
Size: 94349 bytes
Desc: image005.png
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20250415/9fc46e85/attachment-0005.png>