Lingolol

your auntifa liza 🇵🇷 🦛 🦦RE: <a href="https://mastodon.social/@SouthDakota/115520508392114834" rel="nofollow noopener" translate="no" target="_blank">https://mastodon.social/@SouthDakota/115520508392114834</a>okeydockey, since we now have embedded toots here’s an <a href="https://mastodon.social/tags/AltTxt" class="mention hashtag" rel="nofollow noopener" target="_blank">#AltTxt</a> <a href="https://mastodon.social/tags/yaddaYadda" class="mention hashtag" rel="nofollow noopener" target="_blank">#yaddaYadda</a> because it seems <a href="https://mastodon.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> doesn’t exist in South Dakota… AltTxt: there is a masked man with a green face-features-altering mask and a backwards baseball cap inside a car. he is pointing a gun at the photographer, like in a drive-by shooting. ironically, the masked guy looks ambiguously brown-skinned. the caption: "Your tax dollars at work in Chicago yesterday."(implying the gang banger is working for Trump’s <a href="https://mastodon.social/tags/DHS" class="mention hashtag" rel="nofollow noopener" target="_blank">#DHS</a> <a href="https://mastodon.social/tags/ICE" class="mention hashtag" rel="nofollow noopener" target="_blank">#ICE</a>)

Oliver AmmannIs there a viable <a href="https://swiss.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> solution for <a href="https://swiss.social/tags/ancientGreek" class="mention hashtag" rel="nofollow noopener" target="_blank">#ancientGreek</a> text in 16th century prints? Ideas? Experiences?<a href="https://swiss.social/tags/rarebooks" class="mention hashtag" rel="nofollow noopener" target="_blank">#rarebooks</a> <a href="https://swiss.social/tags/greek" class="mention hashtag" rel="nofollow noopener" target="_blank">#greek</a> <a href="https://swiss.social/tags/tesseract" class="mention hashtag" rel="nofollow noopener" target="_blank">#tesseract</a> <a href="https://swiss.social/tags/transkribus" class="mention hashtag" rel="nofollow noopener" target="_blank">#transkribus</a> <a href="https://swiss.social/tags/16thcentury" class="mention hashtag" rel="nofollow noopener" target="_blank">#16thcentury</a>

athmane mokraoui [BoF] ⏚ꝃ⌁⁂Plaque de rue trilingue dans une commune algérienne.L'<a href="https://mstdn.fr/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> ne fonctionne pas en kabyle.

Jeanne (spellboundblog)When will AI understand document semantics? Examining how LLMs may (or may not?) help improve OCR of text - with examples of reading across column breaks.<a href="https://pdfa.org/when-will-ai-understand-document-semantics" rel="nofollow noopener" translate="no" target="_blank">https://pdfa.org/when-will-ai-understand-document-semantics</a><a href="https://digipres.club/tags/ipres2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#ipres2025</a> <a href="https://digipres.club/tags/ocr" class="mention hashtag" rel="nofollow noopener" target="_blank">#ocr</a>

flumen_calculiJetzt mal angenommen, ich wollte auf meinem Mac Scans/Fotos (pdf/jpg/png) von handschriftlichen Notizen in Text umwandeln. Es braucht keine großartige Automatisierung und die Texte sind eher kurz.Kennt Ihr da eine App, mit die dafür geeignet ist und die auch mit der vereinfachten Ausgangsschrift zurechtkommt?<a href="https://ruhr.social/tags/boost" class="mention hashtag" rel="nofollow noopener" target="_blank">#boost</a> <a href="https://ruhr.social/tags/followerpower" class="mention hashtag" rel="nofollow noopener" target="_blank">#followerpower</a> <a href="https://ruhr.social/tags/ocr" class="mention hashtag" rel="nofollow noopener" target="_blank">#ocr</a> <a href="https://ruhr.social/tags/macos" class="mention hashtag" rel="nofollow noopener" target="_blank">#macos</a>

Lambert Heller"Cutting-edge Open OCR Models / We’ve seen an incredible wave of new models this past year. Because so much work is happening in the open, these players build on and benefit from each other’s work. A great example is AllenAI’s release of OlmOCR, which not only released a model but also the dataset used to train it. With these, others can build upon them in new directions. The field is incredibly active, but it’s not always obvious which model to use."<a href="https://openbiblio.social/tags/vlm" class="mention hashtag" rel="nofollow noopener" target="_blank">#vlm</a> <a href="https://openbiblio.social/tags/atr" class="mention hashtag" rel="nofollow noopener" target="_blank">#atr</a> <a href="https://openbiblio.social/tags/ocr" class="mention hashtag" rel="nofollow noopener" target="_blank">#ocr</a> <a href="https://toot.cafe/@tomayac/115418110661215543" rel="nofollow noopener" translate="no" target="_blank">https://toot.cafe/@tomayac/115418110661215543</a>

athmane mokraoui [BoF] ⏚ꝃ⌁⁂Je trouve que c'est une « comique » qu'en <a href="https://mstdn.fr/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> ont nous force à choisir une langue.Genre le texte obtenu tu vas l'avoir dans la langue voulue 🤭Si on parle de script ou de situation, ce serait mieux mais delà à parler de langue en sortie, c'est juste, je ne trouve pas de mots, c'est comique.On pourra parler de langue en sortie s'il y a détection de langue or L'OCR classique ne fait pas de détection de langue.

Terence EdenWhat's working:✅ Download JSON of each page from Amazon. ✅ Deobfuscate the SVG "DRM". ✅ Draw each letter on the page with the correct indent, placement, and font (italics, etc).What's mostly working: 🚧 OCR. Tesseract gets most of the text, but some errors.What's not working: ❌ OCR doesn't output italics. ❌ Linebreaks are hardcoded. ❌ Doesn't integrate into the original ePub code - so no chapters etc. ❌ No idea about footnotes, images, etc.<a href="https://mastodon.social/tags/Kindle" class="mention hashtag" rel="nofollow noopener" target="_blank">#Kindle</a> <a href="https://mastodon.social/tags/DRM" class="mention hashtag" rel="nofollow noopener" target="_blank">#DRM</a> <a href="https://mastodon.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a>

Terence EdenI'm using <a href="https://mastodon.social/tags/tesseract" class="mention hashtag" rel="nofollow noopener" target="_blank">#tesseract</a> 5 to do the <a href="https://mastodon.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> locally.It is *mostly* pretty good - because the default font is nice and clear.But it does have some problems with spaces. Ideally I want this to be fully automated, I don't want to manually correct stuff.Any recommendations for a *local* OCR which runs on <a href="https://mastodon.social/tags/Linux" class="mention hashtag" rel="nofollow noopener" target="_blank">#Linux</a>

Oblomov<a href="https://sociale.network/tags/AskFedi" class="mention hashtag" rel="nofollow noopener" target="_blank">#AskFedi</a> is there an <a href="https://sociale.network/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> for <a href="https://sociale.network/tags/music" class="mention hashtag" rel="nofollow noopener" target="_blank">#music</a> notation? Something that can convert scanned sheet music in some standardized music notation format that can be typeset with appropriate programs?

MiaIn which Sak Supple shares his experiments with using LLMs to transcribe 18th and 19th century playbills from the British Library - manicules, long 's' and all! 'blplaybills.org: better search results using LLMs' <a href="https://www.bl.uk/stories/blogs/posts/blplaybills-org-better-search-results-using-llms" rel="nofollow noopener" translate="no" target="_blank">https://www.bl.uk/stories/blogs/posts/blplaybills-org-better-search-results-using-llms</a><a href="https://hcommons.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> <a href="https://hcommons.social/tags/TheaterHistory" class="mention hashtag" rel="nofollow noopener" target="_blank">#TheaterHistory</a> <a href="https://hcommons.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#LLMs</a> <a href="https://hcommons.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#AI</a> <a href="https://hcommons.social/tags/ATR" class="mention hashtag" rel="nofollow noopener" target="_blank">#ATR</a>

your auntifa liza 🇵🇷 🦛 🦦just the top stub for posts without <a href="https://mastodon.social/tags/AltText" class="mention hashtag" rel="nofollow noopener" target="_blank">#AltText</a> that i <a href="https://mastodon.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> because <a href="https://mastodon.social/tags/yaddaYadda" class="mention hashtag" rel="nofollow noopener" target="_blank">#yaddaYadda</a>

Jürgen HubertFrage: Kann jemand gute <a href="https://mementomori.social/tags/FOSS" class="mention hashtag" rel="nofollow noopener" target="_blank">#FOSS</a> <a href="https://mementomori.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> -Werkzeuge empfehlen, die gut mit <a href="https://mementomori.social/tags/Fraktur" class="mention hashtag" rel="nofollow noopener" target="_blank">#Fraktur</a> -Schrift umgehen können?

Alexander Winkler<a href="https://fedihum.org/@sarahalang" class="u-url mention" rel="nofollow noopener" target="_blank">@sarahalang</a> at "Digital Neo-Latin studies: ideas and perspectives" on efficient <a href="https://openbiblio.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> Post-Correction.<a href="https://openbiblio.social/tags/neolatin" class="mention hashtag" rel="nofollow noopener" target="_blank">#neolatin</a>

Ethan Black<a href="https://fandom.ink/@Fragglemuppet" class="u-url mention" rel="nofollow noopener" target="_blank">@Fragglemuppet</a> <a href="https://mastodon.social/@JKrotkov" class="u-url mention" rel="nofollow noopener" target="_blank">@JKrotkov</a> You'd think modern accessibility tech would use <a href="https://fosstodon.org/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> ...

⚯ Michel de Cryptadamus ⚯Released v1.17.0 of The Pdfalyzer, the surprisingly popular tool for analyzing (possibly malicious) PDFs I created after my own unpleasant experience. Now ships with two command line tools for extracting stuff from PDF files:1. extract_text_from_pdfs() - brute force extract all text from a PDF, including doing an <a href="https://universeodon.com/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> extraction of any embedded images2. extract_pdf_pages() - rip a page range from a <a href="https://universeodon.com/tags/PDF" class="mention hashtag" rel="nofollow noopener" target="_blank">#PDF</a> and write them to a new one* Github: <a href="https://github.com/michelcrypt4d4mus/pdfalyzer" rel="nofollow noopener" translate="no" target="_blank">https://github.com/michelcrypt4d4mus/pdfalyzer</a> * Pypi: <a href="https://pypi.org/project/pdfalyzer/" rel="nofollow noopener" translate="no" target="_blank">https://pypi.org/project/pdfalyzer/</a> * Homebrew: <a href="https://formulae.brew.sh/formula/pdfalyzer" rel="nofollow noopener" translate="no" target="_blank">https://formulae.brew.sh/formula/pdfalyzer</a> * Fun thread someone made last week using Pdfalyzer to explain some of how byzantine the PDF format is: <a href="https://x.com/VikParuchuri/status/1965773078585344215" rel="nofollow noopener" translate="no" target="_blank">https://x.com/VikParuchuri/status/1965773078585344215</a><a href="https://universeodon.com/tags/pypi" class="mention hashtag" rel="nofollow noopener" target="_blank">#pypi</a> <a href="https://universeodon.com/tags/python" class="mention hashtag" rel="nofollow noopener" target="_blank">#python</a> <a href="https://universeodon.com/tags/pdf" class="mention hashtag" rel="nofollow noopener" target="_blank">#pdf</a> <a href="https://universeodon.com/tags/pdfs" class="mention hashtag" rel="nofollow noopener" target="_blank">#pdfs</a> <a href="https://universeodon.com/tags/malware" class="mention hashtag" rel="nofollow noopener" target="_blank">#malware</a> <a href="https://universeodon.com/tags/Threatassessment" class="mention hashtag" rel="nofollow noopener" target="_blank">#Threatassessment</a> <a href="https://universeodon.com/tags/maldoc" class="mention hashtag" rel="nofollow noopener" target="_blank">#maldoc</a> <a href="https://universeodon.com/tags/malwareanalysis" class="mention hashtag" rel="nofollow noopener" target="_blank">#malwareanalysis</a> <a href="https://universeodon.com/tags/homebrew" class="mention hashtag" rel="nofollow noopener" target="_blank">#homebrew</a> <a href="https://universeodon.com/tags/infosec" class="mention hashtag" rel="nofollow noopener" target="_blank">#infosec</a> <a href="https://universeodon.com/tags/cybersecurity" class="mention hashtag" rel="nofollow noopener" target="_blank">#cybersecurity</a> <a href="https://universeodon.com/tags/yararule" class="mention hashtag" rel="nofollow noopener" target="_blank">#yararule</a> <a href="https://universeodon.com/tags/PdfFies" class="mention hashtag" rel="nofollow noopener" target="_blank">#PdfFies</a>

CrossAsiaMajor milestone! 121 million Chinese characters from 1,937 historial titles are now fully searchable thanks to OCR collaboration between <a href="https://openbiblio.social/@stabi_berlin" class="u-url mention" rel="nofollow noopener" target="_blank">@stabi_berlin</a> East Asia Department and Academia Sinica Taiwan. Read more: <a href="http://sbb.berlin/x05w2" rel="nofollow noopener" translate="no" target="_blank">http://sbb.berlin/x05w2</a> <a href="https://openbiblio.social/tags/DigitalHumanities" class="mention hashtag" rel="nofollow noopener" target="_blank">#DigitalHumanities</a> <a href="https://openbiblio.social/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> <a href="https://openbiblio.social/tags/OpenAcess" class="mention hashtag" rel="nofollow noopener" target="_blank">#OpenAcess</a> <a href="https://openbiblio.social/tags/CrossAsia" class="mention hashtag" rel="nofollow noopener" target="_blank">#CrossAsia</a>

de.hypothesesStrg + F statt digitales Blättern: <a href="https://troet.cafe/@manuel_kamenzin" class="u-url mention" rel="nofollow noopener" target="_blank">@manuel_kamenzin</a> nimmt die automatische Texterkennung als Ausgangspunkt, um Studierende in die Welt der <a href="https://fedihum.org/tags/DigitalHumanities" class="mention hashtag" rel="nofollow noopener" target="_blank">#DigitalHumanities</a> einzuführen. Ein Bericht aus der Praxis inklusive Übungen, die den Studierenden verdeutlichen, wie sich ihr Arbeitsalltag durch digitale Methoden entscheidend erleichtern lässt – mit überschaubarem Aufwand 👇<a href="https://digitrip.hypotheses.org/3787" rel="nofollow noopener" translate="no" target="_blank">https://digitrip.hypotheses.org/3787</a> via <a href="https://fedihum.org/@digitrip" class="u-url mention" rel="nofollow noopener" target="_blank">@digitrip</a> fedihum.org<a href="https://fedihum.org/tags/Lehre" class="mention hashtag" rel="nofollow noopener" target="_blank">#Lehre</a> <a href="https://fedihum.org/tags/OCR" class="mention hashtag" rel="nofollow noopener" target="_blank">#OCR</a> <a href="https://fedihum.org/tags/WissenschaftlichesArbeiten" class="mention hashtag" rel="nofollow noopener" target="_blank">#WissenschaftlichesArbeiten</a>

Alíz HorváthComing up on October 3-4, 2025 at Central European University in Vienna: OCR/HTR Workshop for Under-resourced and Under-represented Languages in Digital Humanities, funded by the Cluster of Excellence EurAsian Transformations and by CLARIAH-AT! (Main organizer: yours truly) <a href="https://fedihum.org/tags/digitalhumanities" class="mention hashtag" rel="nofollow noopener" target="_blank">#digitalhumanities</a> <a href="https://fedihum.org/tags/multilingualdh" class="mention hashtag" rel="nofollow noopener" target="_blank">#multilingualdh</a> <a href="https://fedihum.org/tags/textrecognition" class="mention hashtag" rel="nofollow noopener" target="_blank">#textrecognition</a> <a href="https://fedihum.org/tags/ocr" class="mention hashtag" rel="nofollow noopener" target="_blank">#ocr</a> <a href="https://fedihum.org/tags/htr" class="mention hashtag" rel="nofollow noopener" target="_blank">#htr</a>

ZfdGWer heute noch etwas interessantes zu lesen braucht, kann sich diesen frisch bei uns erschienenen Fachartikel zu Gemüte führen:Norbert Fischer, Dominik Kimmel und Frank Puppe berichten über ein Experiment, bei dem die Texte von beschrifteten Bildkarten des Leibniz-Zentrums für Archäologie (LEIZA) einerseits mit einer klassischen <a href="https://fedihum.org/tags/Deep" class="mention hashtag" rel="nofollow noopener" target="_blank">#Deep</a>-Learning-Pipeline und andererseits mit <a href="https://fedihum.org/tags/Large" class="mention hashtag" rel="nofollow noopener" target="_blank">#Large</a>-Language-Modellen (LLMs) erschlossen werden: <a href="https://doi.org/10.17175/2025_09" rel="nofollow noopener" translate="no" target="_blank">https://doi.org/10.17175/2025_09</a><a href="https://fedihum.org/tags/DigitalHumanities" class="mention hashtag" rel="nofollow noopener" target="_blank">#DigitalHumanities</a> <a href="https://fedihum.org/tags/dh" class="mention hashtag" rel="nofollow noopener" target="_blank">#dh</a> <a href="https://fedihum.org/tags/archive" class="mention hashtag" rel="nofollow noopener" target="_blank">#archive</a> <a href="https://fedihum.org/tags/llm" class="mention hashtag" rel="nofollow noopener" target="_blank">#llm</a> <a href="https://fedihum.org/tags/ocr" class="mention hashtag" rel="nofollow noopener" target="_blank">#ocr</a>

Recent searches

Search options

Administered by:

Server stats:

#ocr