Elamur Research⌕
☰
home / papers / 2026 / HAnnoI: A Handwriting Annotation Interface to Extract Data for Linguistic Analyses of Graphetic Detail

HAnnoI: A Handwriting Annotation Interface to Extract Data for Linguistic Analyses of Graphetic Detail

Paper · 2026 · Joshua Wieler, Simon Petitjean, Kristian Berg et al.

Summary

The authors address a practical gap in grapholinguistic research. Scholars studying the visual form of handwritten letters often lack software that lets them mark small regions in scanned pages, attach rich multi-layer annotations to each region, and export the results for statistical work. Many still rely on paper tally sheets. Existing image labelers mostly allow only one label per region or are aimed at producing machine-learning training data.

The paper introduces HAnnoI, an open-source Python/PyQt6 desktop tool. Users draw rectangular Regions of Interest (ROIs) on images or PDF pages and can attach an optional baseline "anchor" to each one. Annotations go on user-defined layers. The tool exports everything as a re-importable CSV, along with individual PNG crops linked to the CSV by a shared index.

To demonstrate it, five annotators marked lower-case <s> in handwritten texts from the GraphVar Corpus. An exploratory analysis then looked at what predicts cursive versus block-letter <s>. Preceding letters appeared to matter more than following ones. HAnnoI thus supports reproducible, extensible workflows for studying graphetic variation.

Key points

  • HAnnoI is written in Python with PyQt6, runs from a single script, and is freely available at github.com/pywielR/HAnnoI.
  • Each ROI stores:
    • its source document and page;
    • its index;
    • its coordinates (x, y, width, height);
    • an optional anchor;
    • any number of custom annotation layers.
  • Work is saved by exporting to CSV, which can be re-imported later, and per-ROI PNG files can also be extracted.
  • The authors compare HAnnoI with VIA, IMMARKUS, eScriptorium, makesense.ai, PixLab Annotate and UAM ImageTool. They consider IMMARKUS the closest alternative, while VIA offers less precise marking and no image extraction.
  • Current limitations:
    • ROIs can only be rectangles.
    • Annotation layers are text fields only, plus a categorical annotation mode, with no drop-down menus or checkboxes.
    • The anchor design favors horizontal, left-to-right scripts.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Heinrich Heine University DüsseldorfDüsseldorf · Germany34 works · 1,487 citations🎓Ruhr University BochumBochum · Germany27 works · 1,189 citations🎓Carl von Ossietzky Universität OldenburgOldenburg · Germany2 works

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 worksLiterature and Literary Theory · 541 works

Cite

Cite as Joshua Wieler, Simon Petitjean, Kristian Berg, Henriette Huber and Stefan Hartmann. 2026. HAnnoI: A Handwriting Annotation Interface to Extract Data for Linguistic Analyses of Graphetic Detail. Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/2i7h8bznd292

@inproceedings{wieler2026hannoi,
  title = {{HAnnoI: A Handwriting Annotation Interface to Extract Data for Linguistic Analyses of Graphetic Detail}},
  author = {Joshua Wieler, Simon Petitjean, Kristian Berg, Henriette Huber and Stefan Hartmann},
  booktitle = {Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026},
  year = {2026},
  doi = {10.63317/2i7h8bznd292},
  url = {https://doi.org/10.63317/2i7h8bznd292}
}

References

8 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Kristian Berg, Stefan Hartmann, and Daniel Claeser. 2023. Are some morphological units more prone to spelling variation than others? a case study using spontaneous handwritten data. Morphology, 34. find
  2. Hiroki Nakayama, Takahiro Kubo, Junya Kamura, Yasufumi Taniguchi, and Xu Liang. 2018. doccano: Text annotation tool for human. Software available from https://github.com/doccano/doccano. ↗
  3. Kristian Berg, Jonas Romstadt, and Cedrek Neitzert. 2021. Graphvar–korpusaufbau und annotation version 1.0. Online-Dokumentation: https://graphvar.uni-bonn. de/dokumentation. ↗
  4. Simon Petitjean, Kristian Berg, Joshua Wieler, Henriette Huber, and Stefan Hartmann. n.d. Are handwritten letter forms shaped by morphology? Manuscript in preparation. find
  5. Hilde De Weerdt, Rainer Simon, Dawn Zhuang, Sunkyu Lee, Iva Stojević, Meret Meister, and Xi Wangzhi. 2024. IMMARKUS: Image annotation in X-MARKUS. Web application. find
  6. Niklas Reinken. 2023. Die Grammatik der Handschriften. find
  7. Piotr Skalski. 2019. Make Sense. https://github.com/SkalskiP/make-sense/. ↗
  8. Han Sloetjes and Peter Wittenburg. 2008. Annotation by category – elan and iso dcr. In Proceedings of the 6th International Conference on Language Resources and Evaluation (LREC 2008), Nijmegen. Cited via ELAN software page. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/wieler-2026-hannoi-handwriting-annotation-interface-extract/