Open-source Chronicles-OCR benchmark evaluates visual perception of ancient Chinese characters
Original title:🎉 🎉 🎉 We're open-sourcing Chronicles-OCR, a visual perception benchmark evaluating VLLMs on ancie...
AI summary
An open-source benchmark, Chronicles-OCR, has been released to evaluate the visual perception capabilities of vision large language models (VLLM) on ancient Chinese characters. The dataset covers 3,000 years of evolution from oracle bone script to cursive script, including 7 historical scripts and 2,800 balanced images. The evaluation covers four core tasks: glyph localization, fine-grained recognition, ancient text parsing, and font classification, aiming to explore how visual distribution changes over time affect model perception. The related paper and code have been open-sourced.
Key Takeaways
- An open-source benchmark,Chronicles-OCR,has been released to evaluate the visual perception…
- The dataset covers 3,000 years of evolution from oracle bone script to cursive script,including 7…
- The evaluation covers four core tasks: glyph localization,fine-grained recognition,ancient text…
- The related paper and code have been open-sourced.
Quick Facts
- Published
- Updated
- Source
- X:腾讯混元 (@TencentHunyuan)
- Category
- Papers & Research
- AI score
- 72
腾讯混元开源的视觉感知基准,专攻古汉字识别,覆盖从甲骨文到草书的三千年演变,做 OCR 和视觉模型的可以拿来测测自家模型在历史文本上的感知退化。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Open-source Chronicles-OCR benchmark evaluates visual perception of ancient Chinese characters. May 19, 2026. https://new.nat6.net/en/news/cmpc1uga61c67slnzs7cijp9e/