Add Arabic model to kraken OCR
Open, Needs TriagePublicFeature
Actions

Assigned To

Authored By

	hubaishan
	Aug 31 2025, 6:09 PM

Project Tags

Referenced Files

None

Subscribers

Description

Wikimedia kraken OCR now does not have any Arabic models, Please add these Arabic models:

		Status	Subtype	Assigned	Task
		Open	Feature	None	T345055 Add kraken OCR engine to Wikimedia OCR
		Open	Feature	Dumbledore	T403346 Add Arabic model to kraken OCR

Restricted Application added subscribers: Gerges, alaa, Aklapper. · View Herald Transcript

I would like to take the task. I have few questions regarding this task. I have downloaded the repo from (https://github.com/wikimedia/wikimedia-ocr)

Kraken isn't integrated yet — should implementing the full Kraken engine (not just the Arabic models) be in scope here?

Should both Zenodo models (all_arabic_scripts.mlmodel + arabic_best.mlmodel) be added, or just the Arabic-specific one?

Thanks!

Can you just clarify the above queries regarding this task

Both models are now installed on https://kraken-ocr.wmcloud.org/. Maybe someone who can read Arabic can test it.

Thank you @sweil
Unfortunately, the results weren't what we hoped for, even after trying well-printed texts.
Example1, Example2