---
title: 7-Scenes
canonical_url: "https://www.modelscope.cn/datasets/OmniData/7-Scenes"
md_url: "https://www.modelscope.cn/datasets/OmniData/7-Scenes.md"
repository: OmniData/7-Scenes
last_updated: 2024-07-14
license: "[7-Scenes Custom]"
storage_size: "20 GB"
domain:
  - publishDate
  - taskTypes
  - displayName
  - publishUrl
  - paperUrl
  - mediaTypes
  - publisher
tasks:
  - 2013-01-01
  - "3D Reconstruction"
  - 7-Scenes
  - "https://www.microsoft.com/en-us/research/project/rgb-d-dataset-7-scenes/"
  - "https://www.microsoft.com/en-us/research/wp-content/uploads/2013/10/glocker2013ismar.pdf"
  - Image
  - Microsoft
downloads: 2183
stars: 0
---

# 7-Scenes

> 7-Scenes - OmniData 在 ModelScope 开源的数据集。displayName: 7-Scenes license: 7-Scenes Custom mediaTypes: Image paperUrl: https://www.microsoft.com/en-us/research/wp-content/uploads/2013/10/glocker2013ismar.pdf publishDate: "2013-01-01" publishUrl:…

OmniData/7-Scenes 是 ModelScope 魔搭社区上的2013-01-01、3D Reconstruction、7-Scenes数据集，涉及 publishDate、taskTypes、displayName 领域，存储大小 20 GB，采用 [7-Scenes Custom] 许可。

- **Repository**: OmniData/7-Scenes
- **License**: [7-Scenes Custom]
- **Tasks**: 2013-01-01, 3D Reconstruction, 7-Scenes, https://www.microsoft.com/en-us/research/project/rgb-d-dataset-7-scenes/, https://www.microsoft.com/en-us/research/wp-content/uploads/2013/10/glocker2013ismar.pdf, Image, Microsoft
- **Domain**: publishDate, taskTypes, displayName, publishUrl, paperUrl, mediaTypes, publisher
- **Storage size**: 20 GB
- **Downloads**: 2183
- **Stars**: 0
- **Last updated**: 2024-07-14

Source: https://www.modelscope.cn/datasets/OmniData/7-Scenes

---

displayName: 7-Scenes
license:
- 7-Scenes Custom
mediaTypes:
- Image
paperUrl: https://www.microsoft.com/en-us/research/wp-content/uploads/2013/10/glocker2013ismar.pdf
publishDate: "2013-01-01"
publishUrl: https://www.microsoft.com/en-us/research/project/rgb-d-dataset-7-scenes/
publisher:
- Microsoft
tags:
- Ground Truth
taskTypes:
- 3D Reconstruction

---
# 数据集介绍
  ## 简介
  7-Scenes 数据集是跟踪的 RGB-D 相机帧的集合。该数据集可用于评估不同应用的方法，例如密集跟踪和映射以及重新定位技术。所有场景均由手持式 Kinect RGB-D 相机以 640×480 分辨率记录。数据集创建者使用 KinectFusion 系统的实现来获取“地面实况”相机轨迹和密集的 3D 模型。每个场景由不同的用户记录几个序列，并分成不同的训练和测试序列集。
  
## Download dataset
:modelscope-code[]{type="git"}