The Learnable Typewriter: A Generative Approach to Text Line Analysis
-
Updated
Oct 31, 2024 - Python
The Learnable Typewriter: A Generative Approach to Text Line Analysis
This is the official implementation of the architecture presented in Leveraging Morphology for Historical Script Metrological Analysis (ICDAR 2026)
Code for the analysis and experiments in the paper “Leveraging Morphology for Historical Script Metrological Analysis”, accepted at ICDAR 2026.
Modular tool for Digital Humanities: IIIF downloader + Studio environment. Supports PDF import, hybrid OCR/HTR, side-by-side manual correction, and global library search. 🛠️📜
DH Scraping and Analyzing the ASC database with Jupyter Notebooks
Library of cu and liturgics resources hosted by Ponomar
Empirical, visual, and computational research on the Voynich Manuscript (Beinecke MS 408)
基于Rime的契丹小字五笔输入方案 | A Rime-based Wubi input schema for the Khitan Small Script
字形大圖譜 (字形大図譜) · A dataset of character forms in pre-modern books of the Sinosphere: bounding boxes, transcriptions, code points, 字母 and variants, CC BY-SA 4.0 · Let's 集字!
古籍竖排 OCR · 人工校对 · 模型训练反哺一体化桌面工具,因命迹(metafuyou.com)而生 | Ancient Chinese vertical-text OCR with human-in-the-loop correction & training
基于Rime的契丹大字五笔输入方案 | A Rime-based Wubi input schema for the Khitan Large Script
Character-level OCR for Old Church Slavonic manuscripts using YOLOv8, synthetic data, reading-order reconstruction, and ONNX/Flutter deployment.
API de l’application Adele (édition diplomatique collaborative)
基于Rime的女真文五笔输入方案 | A Rime-based Wubi input schema for the Jurchen Script
An open, research-focused tool for the analysis of Papyrus 46
Asistente experimental de transcripción paleográfica con IA multimodal para manuscritos hispánicos de los siglos XV–XVIII.
Type anything and drag time back 4,000 years — every letter of the alphabet morphs through its real historical forms into the Egyptian picture it began as. One HTML file.
Image of a civil-registry / parish record to validated structured JSON. BYOK, provider-agnostic, no embedded model, no storage.
Machine-learning palaeography of the Voynich manuscript: do Lisa Fagin Davis's five scribes show in 7,096 handwritten letters k embedded with DINOv3?
AKT Reader - Evidence Lab: reproducible evaluations, labels, validators, and research utilities; not the reader app or benchmark.
To associate your repository with the paleography topic, visit your repo's landing page and select "manage topics."