Thrilled to share the release of this exciting new project! Designed to be enjoyed by visitors of any age, come check out these amazing works of art and find fun activities for yourself or your tiny humans! Want to learn about tarashikomi? We got you! Want to color a picture of a carp? You can! 🐟
For Japanese Studies folks: I’ve ported the NDL’s kuzushiji OCR to run directly in your web browser.
ndlkotenocr-lite-web.netlify.app
No GPU or installation needed. First-time use may take a few minutes to download the model, but after that, it should transcribe each image in about ten seconds.
The system runs directly in your web browser and can recognize ruby annotations and kunten marks in kanbun. Please give it a try and let me know your feedback!
Japanese studies folks,
I’m pleased to introduce Minna de Honkoku OCR, a new AI system for recognizing kuzushiji.
yuta1984.github.io/honkoku-ocr-...
It is an AI-OCR system trained on 57 million characters of input data from the crowdsourced transcription platform Minna de Honkoku.
📢 Digital humanities & Japanese studies researchers,
I’ve developed a JavaScript library for layout analysis of classical Japanese books 📚
👉 GitHub: github.com/yuta1984/kot...
🚀 Live demo: koten-layout.netlify.app
