#iconclass
@cthoyt what comes to mind:
CIDOC CRM, Getty Vocabularies, Europeana Data Model (EDM), Iconclass
#digitalhumanities #DigitalArtHistory #ontology #ai
November 4, 2025 at 4:44 PM
Drew B. Thomas: Automating Iconclass: LLMs and RAG for Large-Scale Classification of Religious Woodcuts https://arxiv.org/abs/2510.19986 https://arxiv.org/pdf/2510.19986 https://arxiv.org/html/2510.19986
October 24, 2025 at 6:32 AM
Drew B. Thomas
Automating Iconclass: LLMs and RAG for Large-Scale Classification of Religious Woodcuts
https://arxiv.org/abs/2510.19986
October 24, 2025 at 5:41 AM
[2025-10-24] 📚 Updates in #MLLM

(1) <a href="https://researchtrend.ai/papers/2510.19986" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link" target="_blank" rel="noopener" data-link="bsky">Automating Iconclass: LLMs and RAG for Large-Scale Classification of Religious Woodcuts
(2) Automating Iconclass: LLMs and RAG for Large-Scale Classification of Religious Woodcuts

🔍 More at researchtrend.ai/communities/MLLM
October 24, 2025 at 3:10 AM
@nfdi4ing I see you there. Have any ideas on how we can do something in your space? E.g., what kind of engineering items do you think might appear in Iconclass that we could connect to your KG?

is https://ingest.nfdi4ing.de/api/v1/sparql a good place to start? Is there a front-end for this / docs?
October 10, 2025 at 10:25 AM
@felwert @danielvanstrien over at LinkedId, Daniel wrote "This is an early release - feedback from GLAM professionals very welcome!"

Oh, and I also found the blog post with more technical details: https://danielvanstrien.xyz/posts/2025/iconclass-vlm-sft/trl-vlm-fine-tuning-iconclass.html
Fine-tuning VLMs for Art History with TRL and HF Jobs
Train vision-language models to generate Iconclass metadata for artworks using TRL’s VLM support and cloud GPUs - no local setup needed!
danielvanstrien.xyz
September 9, 2025 at 8:08 AM
Nice! HuggingFace's Daniel van Strien (@danielvanstrien) has released a Qwen2.5-based vision-language model to generate #iconclass classification codes from images:
https://huggingface.co/davanstrien/iconclass-vlm

Test space: https://huggingface.co/spaces/davanstrien/iconclass-predictions
davanstrien/iconclass-vlm · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
September 9, 2025 at 6:36 AM
Blogged: Fine-tuning a VLM for art history in hours, not weeks

iconclass-vlm generates museum catalog codes (fun fact: "71H7131" = "Bathsheba with David's letter"!)

@hf.co TRL + Jobs = magic ✨

Guide here: danielvanstrien.xyz/posts/2025/i...
danielvanstrien.xyz
September 4, 2025 at 8:19 PM
🌊✨ 3-night cruise from €1150pp ✨🌊
🚢Onboard the incredible #LegendoftheSeas!
Be part of history on the final European sailing of Royal Caribbean’s #IconClass before it returns to the Caribbean 🏝️
jwttravel.ie/tours/rome-t...
#RoyalCaribbean #JWTCruises #TravelGoals #CruiseOffers
September 4, 2025 at 10:07 AM
I fine-tuned a smol VLM to generate specialized art history metadata!

iconclass-vlm: Qwen2.5-VL-3B trained using SFT to generate ICONCLASS codes (think Dewey Decimal for art!)

Trained with @hf.co TRL + Jobs - single UV script, no GPU needed!

Blog soon!
September 3, 2025 at 6:22 PM
I had a great time in Saint-Étienne recently, speaking about multi-modal computer vision and Iconclass. Got to see a bit of Lyon too.
June 30, 2025 at 4:46 PM
Curious about ICONCLASS?

👉 You can still register for the two-part online workshop “Iconclass - Content Classification and Technical Features” on Monday, June 30 and Tuesday, July 1 (10:00 - 13:00)

nfdi4culture.de/events/detai...
Two Part Workshop "Iconclass – Content Classification and Technical Features” - NFDI4Culture
nfdi4culture.de
June 17, 2025 at 1:42 PM
Are often better served by specialist classification schemes such as Iconclass, Dickinson, or NLM classification.
In fact I think the same would be true of food, with the Chewy Decimal working for most grocery stores and a different system better for corner shops
June 15, 2025 at 7:32 AM
Not only Iconclass, but the KG(s) in which it gets applied.
I am building graph embeddings for RDF.

My main use-case is an RDF KG with circa 80 million triples, and it will probably grow more in the coming year.
May 2, 2025 at 12:17 PM
Ahh hmm yeah that would be problematic. If you already have a Python backend and a GPU then cuGraph may work really well for you. Is this for Iconclass?
May 2, 2025 at 11:29 AM
In cooperation with the data competence centre HERMES, NFDI4Culture is hosting a two part online workshop on "Iconclass – Content Classification and Technical Features".

nfdi4culture.de/events/save-...

Save the date, 30 June & 1 July 2025
Save the Date: Two Part Iconclass Workshop - NFDI4Culture
nfdi4culture.de
April 22, 2025 at 10:07 AM
Fab! Thanks for putting in the work.
Now I wish I had the time to put in some effort on doing the same for the Iconclass testset.
April 1, 2025 at 9:49 AM
Baybeeeeeeeeeee, that Masked Singer Finale was a Musicality Iconclass!! *chef's kiss* BOTH were winners for me. Justa Sangin!!!
a man in a yellow dress singing into a microphone with the letter n in the corner
ALT: a man in a yellow dress singing into a microphone with the letter n in the corner
media.tenor.com
December 19, 2024 at 2:09 AM
A fascinating presentation by Ralph Ewert, Leibniz University Hannover: Unlocking Cultural Heritage: Computer Vision for Art and History Archives with the platform iART with semantic search, geoestimation, iconclass, pose estimation, date estimation www.iart.vision
iART
An Interactive Analysis and Retrieval Tool to Support Image-Oriented Research Processes
www.iart.vision
December 12, 2024 at 4:34 PM