Research map / Vision-language and multimodal / Human and embodied modeling
MORE: A Multilingual Document Parsing Benchmark and Evaluation
Long Xu, Binghong Wu, TingHao YU, Hao Feng, Zhenyu Huang, Haoqing Jiang, Yunhao Wang, Shuo Huang, feng zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use a specialized document-parsing VLM (HunyuanOCR) instead of general VLM Qwen3-VL | HunyuanOCR · MORE overall | Overall score: 92.42 vs 83.56 | Improved |
| use HunyuanOCR instead of PaddleOCR-VL | HunyuanOCR · MORE table recognition | TEDS: 78.56 vs 61.11 | Improved |
| compare dots.ocr table score to its text score | dots.ocr · MORE | Text vs Table: 94.45 -> 39.81 | Mixed |
| score aggregated page-wise instead of task-wise | MinerU2.5 · MORE overall | Overall score: 48.85 vs 51.84 | Mixed |
| score aggregated page-wise instead of task-wise | PaddleOCR-VL · MORE table recognition | TEDS: 61.11 -> 57.15 | Hurt |
| use Qwen3-VL-2B instead of PaddleOCR-VL on Sanskrit | Qwen3-VL-2B · MORE Sanskrit overall | Overall score: 70.69 vs 63.27 | Improved |
| evaluate Sanskrit script instead of Latin script for text recognition | HunyuanOCR · MORE text by script | Text NED: 97.51 -> 63.10 | Hurt |
| evaluate Other scripts instead of Latin overall | PaddleOCR-VL · MORE overall by script | Overall score: 93.02 -> 60.06 | Hurt |
| use Qwen2.5-VL-3B instead of PaddleOCR-VL for table recognition on top languages | Qwen2.5-VL-3B · MORE table top-15 languages | TEDS avg: 76.61 vs 67.29 | Improved |
| compare language identification accuracy with text NED | MinerU2.5 · MORE text | Text NED vs language ACC: 27.12 NED, 26.16% ACC | Mixed |
Closest papers by meaning
- MDPBench: A Benchmark for Multilingual Document Parsing in Real-World Scenarios (NeurIPS 2026 · similarity 0.93)
- MMTEB: Massive Multilingual Text Embedding Benchmark (ICLR 2025 · similarity 0.87)
- DocPTBench: Benchmarking End-to-End Photographed Document Parsing and Translation (NeurIPS 2026 · similarity 0.87)
- MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsing (NeurIPS 2026 · similarity 0.87)
- DCAD-2000: A Multilingual Dataset across 2000+ Languages with Data Cleaning as Anomaly Detection (NeurIPS 2025 · similarity 0.86)
- MosaicDoc: A Large-Scale Bilingual Benchmark for Visually Rich Document Understanding (AAAI 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
