Research map / Vision-language and multimodal / CLIP and zero-shot
LaTeX2Layout: High-Fidelity, Scalable Document Layout Annotation Pipeline for Layout Detection
Feijiang Han, Zelong Wang, Bowen Wang, Xinxin Liu, Skyler Cheung, Delip Rao, Chris Callison-Burch, Lyle Ungar
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 |
|---|---|---|---|
| Fine-tune Qwen2.5-VL-3B on LaTeX2Layout data vs zero-shot GPT-4o | Qwen2.5-VL-3B · Layout test set | mAP50 / Kendall tau: tau 0.31 -> 0.95; mAP50 0.34 -> 0.91 | Improved |
| Add 120K synthetic pages to 20K real pages (ablation) | Fine-tuned Qwen2.5-VL-3B · Layout test set | element grounding mAP50: 0.78 -> 0.91 | Improved |
| Add 120K synthetic pages to 20K real pages (ablation) | Fine-tuned Qwen2.5-VL-3B · Layout test set | reading order tau: 0.91 -> 0.95 | Improved |
| Add 120K synthetic pages (ablation) | Fine-tuned Qwen2.5-VL-3B · footnote class | mAP50: 0.57 -> 0.87 | Improved |
| Add 120K synthetic pages (ablation) | Fine-tuned Qwen2.5-VL-3B · math class | mAP50: 0.55 -> 0.90 | Improved |
| Train on 20K synthetic instead of 20K real (ablation) | Qwen2.5-VL-3B · Layout test set | reading order tau: 0.91 -> 0.90 | No clear effect |
| Scale synthetic data from 20K upward (ablation) | Qwen2.5-VL-3B · Layout test set | grounding mAP50 | Improved |
| Scale synthetic data from 20K upward (ablation) | Qwen2.5-VL-3B · Layout test set | reading order tau | No clear effect |
| Add easy-to-hard curriculum re-ranking (ablation) | Qwen2.5-VL-3B (20K real) · Layout test set | reading order tau: 0.88 -> 0.91 | Improved |
| Add easy-to-hard curriculum re-ranking (ablation) | Qwen2.5-VL-3B (20K real) · Layout test set | average mAP50: 0.72 -> 0.78 | Improved |
| Add easy-to-hard curriculum re-ranking (ablation) | Qwen2.5-VL-3B (20K real) · footnote class | mAP50: 0.28 -> 0.57 | Improved |
| Add curriculum re-ranking (ablation) | Qwen2.5-VL-3B (20K real) · Layout test set, per-class | mAP50: Tit 0.77 -> 0.75; TabCap 0.81 -> 0.78; Head 0.66 -> 0.73 | Mixed |
Closest papers by meaning
- A Simple yet Effective Layout Token in Large Language Models for Document Understanding (CVPR 2025 · similarity 0.89)
- LayoutLLM: Layout Instruction Tuning with Large Language Models for Document Understanding (CVPR 2024 · similarity 0.88)
- M2Doc: A Multi-Modal Fusion Approach for Document Layout Analysis (AAAI 2024 · similarity 0.87)
- Rethinking The Training And Evaluation of Rich-Context Layout-to-Image Generation (NeurIPS 2024 · similarity 0.87)
- PosterO: Structuring Layout Trees to Enable Language Models in Generalized Content-Aware Layout Generation (CVPR 2025 · similarity 0.87)
- Seeing is Improving: Visual Feedback for Iterative Text Layout Refinement (CVPR 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.
