---
title: OmniCity
canonical_url: "https://www.modelscope.cn/datasets/OmniData/OmniCity"
md_url: "https://www.modelscope.cn/datasets/OmniData/OmniCity.md"
repository: OmniData/OmniCity
last_updated: 2024-07-13
license: "[CC BY-NC 4.0]"
storage_size: "14 GB"
domain:
  - publishDate
  - paperUrl
  - publishUrl
  - taskTypes
  - labelTypes
  - displayName
  - publisher
tasks:
  - 2022
  - "https://arxiv.org/abs/2208.00928"
  - "https://city-super.github.io/omnicity/"
  - "Instance Segmentation"
  - InstanceSegMap
  - OmniCity
  - "Sun Yat-Sen University"
downloads: 322
stars: 0
---

# OmniCity

> OmniCity - OmniData 在 ModelScope 开源的数据集。displayName: OmniCity labelTypes: InstanceSegMap license: CC BY-NC 4.0 mediaTypes: [] paperUrl: https://arxiv.org/abs/2208.00928 publishDate: "2022" publishUrl: https://city-super.github.io/omnicity/ publisher: Wuhan…

OmniData/OmniCity 是 ModelScope 魔搭社区上的2022、https://arxiv.org/abs/2208.00928、https://city-super.github.io/omnicity/数据集，涉及 publishDate、paperUrl、publishUrl 领域，存储大小 14 GB，采用 [CC BY-NC 4.0] 许可。

- **Repository**: OmniData/OmniCity
- **License**: [CC BY-NC 4.0]
- **Tasks**: 2022, https://arxiv.org/abs/2208.00928, https://city-super.github.io/omnicity/, Instance Segmentation, InstanceSegMap, OmniCity, Sun Yat-Sen University
- **Domain**: publishDate, paperUrl, publishUrl, taskTypes, labelTypes, displayName, publisher
- **Storage size**: 14 GB
- **Downloads**: 322
- **Stars**: 0
- **Last updated**: 2024-07-13

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

---

displayName: OmniCity
labelTypes:
- InstanceSegMap
license:
- CC BY-NC 4.0
mediaTypes: []
paperUrl: https://arxiv.org/abs/2208.00928
publishDate: "2022"
publishUrl: https://city-super.github.io/omnicity/
publisher:
- Wuhan University
- Shanghai Artificial Intelligence Laboratory
- SenseTime Research
- Chinese University of Hong Kong
- Sun Yat-Sen University
tags:
- Building Instance
- Building Type
taskTypes:
- Instance Segmentation

---
# 数据集介绍
  ## 简介
  OmniCity是一个用于全面化城市视觉理解研究的新数据集。该数据集包含从纽约市的25K个地理位置获得的多视角卫星影像、街景全景以及单视角图像，提供了建筑物底座提取、高度估计、建筑物平面/实例/细粒度分割等任务的像素级标注。
  ## 引文
  @article{li2023omnicity,  
         title={OmniCity: Omnipotent City Understanding with Multi-level and Multi-view Images},  
         author={Li, Weijia and Lai, Yawen and Xu, Linning and Xiangli, Yuanbo and Yu, Jinhua and He, Conghui and Xia, Gui-Song and Lin, Dahua},  
         journal={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},  
         year={2023}
 }
  
## Download dataset
:modelscope-code[]{type="git"}