[INDOLOGY] OCR

[Tyler Neill]

Dear List members,

The new drag-and-drop interface to Google Vision OCR that I mentioned last
month is now ready for use on Skrutable. Go straight to the new subpage
skrutable.info/ocr <https://skrutable.info/ocr>, or look for the small link
on the main page, lower-left. The FAQs should answer most questions and get
you up and running in a few minutes.

Many thanks to those who provided detailed feedback! (Arushi, Don, Herman,
Jan, Patrick, Vyom—apologies if others are slipping my mind today.) It
helped me equip the tool with a number of usability features and provide
detailed instructions, especially to hopefully de-scarify Google Billing.
That said, I'll happily make further improvements as needed.

Finally, I learned the great news last week that Dharmamitra may also soon
release a very similar drag-and-drop interface for OCR with Google Gemini,
with no billing setup needed. In my recent tests, Gemini and Cloud Vision
each produce strong results, but they make different errors, suggesting
that combining their outputs could yield the best accuracy. For that use
case, I happen to have another side-project prototype
<https://github.com/tylergneill/squinter> that could prove useful, which
I'll keep tinkering on.

Here for any and all questions, of course!

Kind wishes,
Tyler

On Mon, May 12, 2025 at 3:02 PM Tyler Neill <tyler.g.neill at gmail.com> wrote:

> Hi all,
>
> Regarding Patrick’s question about easy OCR, I suspect he’s particularly
> looking for a tool that can handle multi-page PDFs in one go, which could
> be especially helpful for digitization projects like UTA’s Resource
> Library for Dharmaśāstra Studies
> <https://sites.utexas.edu/sanskrit/resources/dharmasastra/>.
>
> If Patrick or anyone else is interested, feel free to reach out to me
> directly. I’m looking for a few volunteers to test a new drag-and-drop
> interface I’m building to streamline access to Google Vision OCR, which is
> currently best in class and handles multi-page inputs well.
>
> Kind regards,
> Tyler
>
> On Sat, May 10, 2025 at 8:00 AM <indology-request at list.indology.info>
> wrote:
>
>> ---------- Forwarded message ----------
>> From: Patrick Olivelle <jpo at austin.utexas.edu>
>> To: Indology <indology at list.indology.info>
>> Cc:
>> Bcc:
>> Date: Fri, 9 May 2025 22:07:05 +0000
>> Subject: [INDOLOGY] OCR
>> Dear Friends:
>>
>> I am wondering whether with the advance of AI technology we have easy OCR
>> software to read Devanāgarī, easy enough to be used by someone like me!! We
>> have the one prepared by Andrew Ollett, which he generously gave us. But
>> that requires computer knowledge far beyond my reach. Is there on where you
>> can just drop the Devanāgari scan, out pops a searchable file. This is
>> probably a long shot, but I thought I would ask.
>>
>> With thanks and best wishes,
>>
>> Patrick Olivelle
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://list.indology.info/pipermail/indology/attachments/20250616/94d976c7/attachment.htm>