---
title: FUNSD
canonical_url: "https://www.modelscope.cn/datasets/OmniData/FUNSD"
md_url: "https://www.modelscope.cn/datasets/OmniData/FUNSD.md"
repository: OmniData/FUNSD
last_updated: 2024-07-01
license: "[FUNSD Custom]"
storage_size: "17 MB"
domain:
  - publishDate
  - displayName
  - paperUrl
  - publishUrl
  - mediaTypes
  - publisher
  - labelTypes
  - taskTypes
tasks:
  - 2019
  - "FUNSD (Form Understanding in Noisy Scanned Documents)"
  - "https://arxiv.org/pdf/1905.13538v2.pdf"
  - "https://guillaumejaume.github.io/FUNSD/"
  - Image
  - "Istanbul Technical University"
  - Text
  - "Text Relation Extraction"
downloads: 276
stars: 0
---

# FUNSD

> FUNSD - OmniData 在 ModelScope 开源的数据集。displayName: FUNSD (Form Understanding in Noisy Scanned Documents) labelTypes: Semanticsegmap Text license: FUNSD Custom mediaTypes: Image paperUrl: https://arxiv.org/pdf/1905.13538v2.pdf publishDate: "2019" publishUrl:…

OmniData/FUNSD 是 ModelScope 魔搭社区上的2019、FUNSD (Form Understanding in Noisy Scanned Documents)、https://arxiv.org/pdf/1905.13538v2.pdf数据集，涉及 publishDate、displayName、paperUrl 领域，存储大小 17 MB，采用 [FUNSD Custom] 许可。

- **Repository**: OmniData/FUNSD
- **License**: [FUNSD Custom]
- **Tasks**: 2019, FUNSD (Form Understanding in Noisy Scanned Documents), https://arxiv.org/pdf/1905.13538v2.pdf, https://guillaumejaume.github.io/FUNSD/, Image, Istanbul Technical University, Text, Text Relation Extraction
- **Domain**: publishDate, displayName, paperUrl, publishUrl, mediaTypes, publisher, labelTypes, taskTypes
- **Storage size**: 17 MB
- **Downloads**: 276
- **Stars**: 0
- **Last updated**: 2024-07-01

Source: https://www.modelscope.cn/datasets/OmniData/FUNSD

---

displayName: FUNSD (Form Understanding in Noisy Scanned Documents)
labelTypes:
- Semantic_seg_map
- Text
license:
- FUNSD Custom
mediaTypes:
- Image
paperUrl: https://arxiv.org/pdf/1905.13538v2.pdf
publishDate: "2019"
publishUrl: https://guillaumejaume.github.io/FUNSD/
publisher:
- ETH Zurich
- Istanbul Technical University
tags:
- Documents
- Graph
taskTypes:
- Semantic Segmentation
- Semantic Role Labeling
- Text Information Extraction
- Text Relation Extraction

---
# 数据集介绍
  ## 简介
  嘈杂扫描文档中的表单理解 (FUNSD) 包含 199 个真实的、完全注释的扫描表单。文档嘈杂且外观差异很大，使得表单理解 (FoUn) 成为一项具有挑战性的任务。所提出的数据集可用于各种任务，包括文本检测、光学字符识别、空间布局分析和实体标记/链接。
  ## 引文
  ```
 @inproceedings{jaume2019funsd,
 title={Funsd: A dataset for form understanding in noisy scanned documents},
 author={Jaume, Guillaume and Ekenel, Hazim Kemal and Thiran, Jean-Philippe},
 booktitle={2019 International Conference on Document Analysis and Recognition Workshops (ICDARW)},
 volume={2},
 pages={1--6},
 year={2019},
 organization={IEEE}
 }
 
 ```
  
## Download dataset
:modelscope-code[]{type="git"}