---
title: CUTE_80
canonical_url: "https://www.modelscope.cn/datasets/OmniData/CUTE_80"
md_url: "https://www.modelscope.cn/datasets/OmniData/CUTE_80.md"
repository: OmniData/CUTE_80
last_updated: 2024-07-02
license: "[CUTE 80 Custom]"
storage_size: "88 MB"
domain:
  - publishDate
  - displayName
  - paperUrl
  - publishUrl
  - mediaTypes
  - publisher
  - taskTypes
  - labelTypes
tasks:
  - 2014
  - "CUTE 80"
  - "http://cs-chan.com/doc/ESWA_2014A.pdf"
  - "http://cs-chan.com/downloads_CUTE80_dataset.html"
  - Image
  - "National University of Singapore"
  - OCR
  - Polygon
downloads: 44
stars: 0
---

# CUTE_80

> CUTE_80 - OmniData 在 ModelScope 开源的数据集。displayName: CUTE 80 labelTypes: Polygon license: CUTE 80 Custom mediaTypes: Image paperUrl: http://cs-chan.com/doc/ESWA2014A.pdf publishDate: "2014" publishUrl: http://cs-chan.com/downloadsCUTE80dataset.html publisher:…

OmniData/CUTE_80 是 ModelScope 魔搭社区上的2014、CUTE 80、http://cs-chan.com/doc/ESWA_2014A.pdf数据集，涉及 publishDate、displayName、paperUrl 领域，存储大小 88 MB，采用 [CUTE 80 Custom] 许可。

- **Repository**: OmniData/CUTE_80
- **License**: [CUTE 80 Custom]
- **Tasks**: 2014, CUTE 80, http://cs-chan.com/doc/ESWA_2014A.pdf, http://cs-chan.com/downloads_CUTE80_dataset.html, Image, National University of Singapore, OCR, Polygon
- **Domain**: publishDate, displayName, paperUrl, publishUrl, mediaTypes, publisher, taskTypes, labelTypes
- **Storage size**: 88 MB
- **Downloads**: 44
- **Stars**: 0
- **Last updated**: 2024-07-02

Source: https://www.modelscope.cn/datasets/OmniData/CUTE_80

---

displayName: CUTE 80
labelTypes:
- Polygon
license:
- CUTE 80 Custom
mediaTypes:
- Image
paperUrl: http://cs-chan.com/doc/ESWA_2014A.pdf
publishDate: "2014"
publishUrl: http://cs-chan.com/downloads_CUTE80_dataset.html
publisher:
- University of Malaya
- National University of Singapore
tags:
- Curve text image
taskTypes:
- OCR

---
# 数据集介绍
  ## 简介
  我们介绍了第一个要公开的弯曲文本数据集，即CUTE80，它由80个复杂背景，透视失真效果和较差分辨率效果 (圆形，S，Z形文本线) 的弯曲文本线图像组成。CUTE80是必需的，以便显示当前文本检测方法在处理弯曲文本方面的能力。这些图像是用数码相机捕获或从Internet检索的室内或室外图像。
  
## Download dataset
:modelscope-code[]{type="git"}