---
title: "Automated German Alt Text Generation for News Charts (2026)"
description: "The authors therefore propose a hybrid pipeline that uses templates for metadata and an MLLM for the descriptive parts. 29 references."
authors: ["Alessia Vannini", "Julia Locher", "Marianne Santaholma", "Claudia Amsler", "Jonas Oesch", "Arzu Coltekin"]
year: 2026
venue: "Proceedings of the 11th Edition of the Swiss Text Analytics Conference"
source: "https://aclanthology.org/2026.swisstext-1.20/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/vannini-2026-automated-german-alt-text-generation/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/vannini-2026-automated-german-alt-text-generation/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/23/) / Automated German Alt Text Generation for News Charts

# Automated German Alt Text Generation for News Charts

**Authors:** Alessia Vannini, Julia Locher, Marianne Santaholma, Claudia Amsler, Jonas Oesch and Arzu Coltekin · **Year:** 2026 · **Venue:** [Proceedings of the 11th Edition of the Swiss Text Analytics Conference](https://md.research.elamur.ai/free/en/venues/swisstext/) ([Edition of the Swiss Text Analytics Conference 2026](https://md.research.elamur.ai/free/en/venues/swisstext/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.swisstext-1.20/) · [PDF](https://aclanthology.org/2026.swisstext-1.20.pdf)

## Summary
News charts are often inaccessible to people with visual impairments (PVIs) because they lack alt texts. Writing these by hand does not fit fast newsroom workflows, and the European Accessibility Act now requires text alternatives. Because earlier automated approaches dealt only with English, this study examines whether a multimodal LLM can write German alt texts for three chart types published by the Neue Zürcher Zeitung (NZZ): line, bar and stacked bar. The texts should follow accessibility guidelines and match human-written references.

The authors designed a four-part alt text schema and refined it with PVIs. They also built what they describe as the first German gold-standard corpus for this task. Alt texts were generated with Gemini 2.5 Flash using few-shot prompting. They were assessed with SBERT similarity, an LLM-as-a-judge (o4-mini) and interviews with PVIs.

The generated texts were rated clear, complete and useful. However, they were longer than the references and less stable in the descriptive sections, especially for complex charts. Metadata output was almost deterministic. The authors therefore propose a hybrid pipeline that uses templates for metadata and an MLLM for the descriptive parts.

## Key points
- **Dataset:** NZZ supplied 168 charts as PNG images with JSON metadata and data values. Of these, 150 are in scope (line, bar, stacked bar). Charts are labelled simple (one data column) or complex (two or more columns).
- **Alt text schema:** each alt text has four parts: metadata, overview, long description and a data table.
- **Gold-standard corpus:** 41 charts in total (12 line, 18 bar, 11 stacked bar). Evaluation used 27 of them, each generated at 3 temperatures × 10 variants, giving 810 outputs.
- **Evaluation:** SBERT semantic similarity, o4-mini as LLM judge, and interviews with five PVIs. The LLM judge and PVI ratings agreed on clarity and completeness, though the LLM was stricter on conciseness and neutrality.
- **Main findings:** metadata was the most robust section. Overviews and long descriptions were longer and less consistent, especially for complex charts. The study is limited by its small participant pool and a single publisher.

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

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works

🧭 Topics (1): [✍️ Text generation](https://md.research.elamur.ai/free/en/topics/text-generation/) · 214 works

## Cite

Cite as: Alessia Vannini, Julia Locher, Marianne Santaholma, Claudia Amsler, Jonas Oesch and Arzu Coltekin. 2026. Automated German Alt Text Generation for News Charts. Proceedings of the 11th Edition of the Swiss Text Analytics Conference.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{vannini2026automated,
  title = {{Automated German Alt Text Generation for News Charts}},
  author = {Alessia Vannini, Julia Locher, Marianne Santaholma, Claudia Amsler, Jonas Oesch and Arzu Coltekin},
  booktitle = {Proceedings of the 11th Edition of the Swiss Text Analytics Conference},
  year = {2026},
  url = {https://aclanthology.org/2026.swisstext-1.20/}
}
```

- [← previous Automated Essay Scoring and Language Certification: Assessing Generalizability, Agreement and Validity for French](https://md.research.elamur.ai/free/en/papers/wilkens-2026-automated-essay-scoring-language-certification/)
- [next → Automated Reformulation of Argumentative Essays to Improve Argument Organization and Development](https://md.research.elamur.ai/free/en/papers/moraes-2026-automated-reformulation-argumentative-essays-improve/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/23/)

## References

29 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. AbilityNet. 2025. An introduction to screen readers. [https://abilitynet.org.uk/factsheets/introduction-screen-readers](https://abilitynet.org.uk/factsheets/introduction-screen-readers). [↗](https://abilitynet.org.uk/factsheets/introduction-screen-readers)
2. Accessibility Guidelines Working Group (AG WG). 2023. Web Content Accessibility Guidelines (WCAG) 2.2 Quick Reference. [https://www.w3.org/WAI/WCAG22/quickref/](https://www.w3.org/WAI/WCAG22/quickref/). [↗](https://www.w3.org/WAI/WCAG22/quickref/)
3. Abhijit Balaji, Thuvaarakkesh Ramanathan, and Venkateshwarlu Sonathi. 2018. Chart-Text: A Fully Automated Chart Image Descriptor. arXiv preprint. [find](https://scholar.google.com/scholar?q=Abhijit+Balaji%2C+Thuvaarakkesh+Ramanathan%2C+and+Venkateshwarlu+Sonathi%2E+2018%2E+Chart%2DText%3A+A+Fully+Automated+Chart+Image+Descriptor%2E+arXiv+preprint%2E)
4. Aspen Belle, Vanessa Goh, Akshay Kumar, Richard Pranjatno, Pui Yip, Umayangani Wickramaratne, and Humphrey Obie. 2022. Alt-Texify: A Pipeline to Generate Alt-text from SVG Visualizations:. In Proceedings of the 17th International Conference on Evaluation of Novel Approaches to Software Engineering, pages 275–281, Online Streaming, — Select a Country —. SCITEPRESS - Science and Technology Publications. [find](https://scholar.google.com/scholar?q=Aspen+Belle%2C+Vanessa+Goh%2C+Akshay+Kumar%2C+Richard+Pranjatno%2C+Pui+Yip%2C+Umayangani+Wickramaratne%2C+and+Humphrey+Obie%2E+2022%2E+Alt%2DTexify%3A+A+Pipeline+to+Generate+Alt%2Dtext+from+SVG+Visualizations%3A%2E+In+Proceedi)
5. Arzu Çöltekin, Amy Griffin, and Anthony Robinson. 2021. Visualizations. Oxford University Press. Consumer Financial Protection Bureau. Data visualization guidelines. [https://cfpb.github.io/design-system/guidelines/data-visualization-guidelines](https://cfpb.github.io/design-system/guidelines/data-visualization-guidelines). [↗](https://cfpb.github.io/design-system/guidelines/data-visualization-guidelines)
6. DIAGRAM Center. 2020. Specific Guidelines - Graphs. [find](https://scholar.google.com/scholar?q=DIAGRAM+Center%2E+2020%2E+Specific+Guidelines+%2D+Graphs%2E)
7. European Union. 2019. Directive (EU) 2019/882 of the European Parliament and of the Council of 17 April 2019 on the accessibility requirements for products and services. [find](https://scholar.google.com/scholar?q=European+Union%2E+2019%2E+Directive+%28EU%29+2019%2F882+of+the+European+Parliament+and+of+the+Council+of+17+April+2019+on+the+accessibility+requirements+for+products+and+services%2E)
8. Ali Mazraeh Farahani, Peyman Adibi, Mohammad Saeed Ehsani, Hans-Peter Hutter, and Alireza Darvishy. 2023. Automatic Chart Understanding: A Review. IEEE Access, 11:76202–76221. [find](https://scholar.google.com/scholar?q=Ali+Mazraeh+Farahani%2C+Peyman+Adibi%2C+Mohammad+Saeed+Ehsani%2C+Hans%2DPeter+Hutter%2C+and+Alireza+Darvishy%2E+2023%2E+Automatic+Chart+Understanding%3A+A+Review%2E+IEEE+Access%2C+11%3A76202%E2%80%9376221%2E)
9. E. Hoque and M. Saidul Islam. 2025. Natural Language Generation for Visualizations: State of the Art, Challenges and Future Directions. Computer Graphics Forum, 44(1):e15266. [find](https://scholar.google.com/scholar?q=E%2E+Hoque+and+M%2E+Saidul+Islam%2E+2025%2E+Natural+Language+Generation+for+Visualizations%3A+State+of+the+Art%2C+Challenges+and+Future+Directions%2E+Computer+Graphics+Forum%2C+44%281%29%3Ae15266%2E)
10. Kung-Hsiang Huang, Hou Pong Chan, May Fung, Haoyi Qiu, Mingyang Zhou, Shafiq Joty, Shih-Fu Chang, and Heng Ji. 2025. From Pixels to Insights: A Survey on Automatic Chart Understanding in the Era of Large Foundation Models. IEEE Transactions on Knowledge and Data Engineering, 37(5):2550– 2568. [find](https://scholar.google.com/scholar?q=Kung%2DHsiang+Huang%2C+Hou+Pong+Chan%2C+May+Fung%2C+Haoyi+Qiu%2C+Mingyang+Zhou%2C+Shafiq+Joty%2C+Shih%2DFu+Chang%2C+and+Heng+Ji%2E+2025%2E+From+Pixels+to+Insights%3A+A+Survey+on+Automatic+Chart+Understanding+in+the+Era+of+La)
11. Crescentia Jung, Shubham Mehta, Atharva Kulkarni, Yuhang Zhao, and Yea-Seul Kim. 2022. Communicating Visualizations without Visuals: Investigation of Visualization Alternative Text for People with Visual Impairments. IEEE Transactions on Visualization and Computer Graphics, 28(1):1095–1105. [find](https://scholar.google.com/scholar?q=Crescentia+Jung%2C+Shubham+Mehta%2C+Atharva+Kulkarni%2C+Yuhang+Zhao%2C+and+Yea%2DSeul+Kim%2E+2022%2E+Communicating+Visualizations+without+Visuals%3A+Investigation+of+Visualization+Alternative+Text+for+People+with+Vis)
12. Shankar Kantharaj, Rixie Tiffany Leong, Xiang Lin, Ahmed Masry, Megh Thakkar, Enamul Hoque, and Shafiq Joty. 2022. [Chart-to-Text: A Large-Scale Benchmark for Chart Summarization](https://md.research.elamur.ai/free/en/papers/kantharaj-2022-chart-text-large-scale-benchmark/). In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4005–4023, Dublin, Ireland. Association for Computational Linguistics.
13. Julia Locher and Alessia Vannini. 2025. Assisting People with Visual Impairments to Understand Data Visualizations. Technical Report 25FS_IIT47, University of Applied Sciences and Arts Northwestern Switzerland (FHNW), School of Computer Science, Windisch, Switzerland. [find](https://scholar.google.com/scholar?q=Julia+Locher+and+Alessia+Vannini%2E+2025%2E+Assisting+People+with+Visual+Impairments+to+Understand+Data+Visualizations%2E+Technical+Report+25FS%5FIIT47%2C+University+of+Applied+Sciences+and+Arts+Northwestern+Sw)
14. Ahmed Masry, Do Long, Jia Qing Tan, Shafiq Joty, and Enamul Hoque. 2022. [ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning](https://md.research.elamur.ai/free/en/papers/masry-2022-chartqa-benchmark-question-answering-about/). In Findings of the Association for Computational Linguistics: ACL 2022, pages 2263–2279, Dublin, Ireland. Association for Computational Linguistics.
15. Cole Gleason, Patrick Carrington, Cameron Cassidy, Meredith Ringel Morris, Kris M. Kitani, and Jeffrey P. Bigham. 2019. “It’s almost like they’re trying to hide it”: How User-Provided Image Descriptions Have Failed to Make Twitter Accessible. In The World Wide Web Conference, pages 549–559, San Francisco CA USA. ACM. [find](https://scholar.google.com/scholar?q=Cole+Gleason%2C+Patrick+Carrington%2C+Cameron+Cassidy%2C+Meredith+Ringel+Morris%2C+Kris+M%2E+Kitani%2C+and+Jeffrey+P%2E+Bigham%2E+2019%2E+%E2%80%9CIt%E2%80%99s+almost+like+they%E2%80%99re+trying+to+hide+it%E2%80%9D%3A+How+User%2DProvided+Image+Descriptio)
16. Omar Moured, Shahid Ali Farooqui, Karin Müller, Shar-ifeh Fadaeijouybari, Thorsten Schwarz, Mohammed Javed, and Rainer Stiefelhagen. 2024. Alt4Blind: A User Interface to Simplify Charts Alt-Text Creation. In Klaus Miesenberger, Petr Peňáz, and Makoto Kobayashi, editors, Computers Helping People with Special Needs, volume 14750, pages 291– 298. Springer Nature Switzerland, Cham. [find](https://scholar.google.com/scholar?q=Omar+Moured%2C+Shahid+Ali+Farooqui%2C+Karin+M%C3%BCller%2C+Shar%2Difeh+Fadaeijouybari%2C+Thorsten+Schwarz%2C+Mohammed+Javed%2C+and+Rainer+Stiefelhagen%2E+2024%2E+Alt4Blind%3A+A+User+Interface+to+Simplify+Charts+Alt%2DText+Creat)
17. Li Gong, Josep Crego, and Jean Senellart. 2019. Enhanced Transformer Model for Data-to-Text Generation. In Proceedings of the 3rd Workshop on Neural Generation and Translation, pages 148–156, Hong Kong. Association for Computational Linguistics. Neue Zürcher Zeitung. NZZ – Neue Zürcher Zeitung. [https://www.nzz.ch](https://www.nzz.ch). [↗](https://www.nzz.ch)
18. Google DeepMind. 2025. Gemini 2.5 Flash (Preview). [find](https://scholar.google.com/scholar?q=Google+DeepMind%2E+2025%2E+Gemini+2%2E5+Flash+%28Preview%29%2E)
19. Kai Nylund, Jennifer Mankoff, and Venkatesh Potluri. 2025. MatplotAlt: A Python Library for Adding Alt Text to Matplotlib Figures in Computational Notebooks. Computer Graphics Forum, 44(3):e70119. [find](https://scholar.google.com/scholar?q=Kai+Nylund%2C+Jennifer+Mankoff%2C+and+Venkatesh+Potluri%2E+2025%2E+MatplotAlt%3A+A+Python+Library+for+Adding+Alt+Text+to+Matplotlib+Figures+in+Computational+Notebooks%2E+Computer+Graphics+Forum%2C+44%283%29%3Ae70119%2E)
20. Jiawei Gu, Xuhui Jiang, Zhichao Shi, Hexiang Tan, Xuehao Zhai, Chengjin Xu, Wei Li, Yinghan Shen, Shengjie Ma, Honghao Liu, Saizhuo Wang, Kun Zhang, Yuanzhuo Wang, Wen Gao, Lionel Ni, and Jian Guo. 2025. A Survey on LLM-as-a-Judge. Preprint, arXiv:2411.15594. [↗](https://arxiv.org/abs/2411.15594)
21. Jason Obeid and Enamul Hoque. 2020. Chart-to-Text: Generating Natural Language Descriptions for Charts by Adapting the Transformer Model. In Proceedings of the 13th International Conference on Natural Language Generation, pages 138–147, Dublin, Ireland. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Jason+Obeid+and+Enamul+Hoque%2E+2020%2E+Chart%2Dto%2DText%3A+Generating+Natural+Language+Descriptions+for+Charts+by+Adapting+the+Transformer+Model%2E+In+Proceedings+of+the+13th+International+Conference+on+Natural)
22. OpenAI. 2024. O4-mini. OpenRouter. OpenRouter API Reference. [https://openrouter.ai/docs/api/reference/parameters](https://openrouter.ai/docs/api/reference/parameters). [↗](https://openrouter.ai/docs/api/reference/parameters)
23. Pennsylvania State University. 2024. Charts & Accessibility. [find](https://scholar.google.com/scholar?q=Pennsylvania+State+University%2E+2024%2E+Charts+%26+Accessibility%2E)
24. Nils Reimers and Iryna Gurevych. 2019. [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://md.research.elamur.ai/free/en/papers/reimers-2019-sentence-bert-sentence-embeddings-siamese/). Preprint, arXiv:1908.10084. Statista. Statista. https://de.statista.com/. [↗](https://arxiv.org/abs/1908.10084)
25. Benny Tang, Angie Boggust, and Arvind Satyanarayan. 2023. VisText: A Benchmark for Semantically Rich Chart Captioning. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7268–7298, Toronto, Canada. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Benny+Tang%2C+Angie+Boggust%2C+and+Arvind+Satyanarayan%2E+2023%2E+VisText%3A+A+Benchmark+for+Semantically+Rich+Chart+Captioning%2E+In+Proceedings+of+the+61st+Annual+Meeting+of+the+Association+for+Computational+Li)
26. Jiaqi Wang, Hanqi Jiang, Yiheng Liu, Chong Ma, Xu Zhang, Yi Pan, Mengyuan Liu, Peiran Gu, Sichen Xia, Wenjun Li, Yutong Zhang, Zihao Wu, Zhengliang Liu, Tianyang Zhong, Bao Ge, Tuo Zhang, Ning Qiang, Xintao Hu, Xi Jiang, and 5 others. 2024. A Comprehensive Review of Multimodal Large Language Models: Performance and Challenges Across Different Tasks. arXiv preprint. [find](https://scholar.google.com/scholar?q=Jiaqi+Wang%2C+Hanqi+Jiang%2C+Yiheng+Liu%2C+Chong+Ma%2C+Xu+Zhang%2C+Yi+Pan%2C+Mengyuan+Liu%2C+Peiran+Gu%2C+Sichen+Xia%2C+Wenjun+Li%2C+Yutong+Zhang%2C+Zihao+Wu%2C+Zhengliang+Liu%2C+Tianyang+Zhong%2C+Bao+Ge%2C+Tuo+Zhang%2C+Ning+Qiang%2C+)
27. WebAIM. 2026. WebAIM: The WebAIM Million. [https://webaim.org/projects/million/](https://webaim.org/projects/million/). [↗](https://webaim.org/projects/million/)
28. Chuqiao Yan, Hans-Peter Hutter, Felix M. Schmitt-Koopmann, and Alireza Darvishy. 2025. Chart Accessibility: A Review of Current Alt Text Generation. IEEE Access, 13:94040–94056. [find](https://scholar.google.com/scholar?q=Chuqiao+Yan%2C+Hans%2DPeter+Hutter%2C+Felix+M%2E+Schmitt%2DKoopmann%2C+and+Alireza+Darvishy%2E+2025%2E+Chart+Accessibility%3A+A+Review+of+Current+Alt+Text+Generation%2E+IEEE+Access%2C+13%3A94040%E2%80%9394056%2E)
29. Shukang Yin, Chaoyou Fu, Sirui Zhao, Ke Li, Xing Sun, Tong Xu, and Enhong Chen. 2024. [A survey on multimodal large language models](https://md.research.elamur.ai/free/en/papers/yin-2023-survey-multimodal-large-language-models/). National Science Review, 11(12):nwae403.

*Updated 2 November 2026*