andreped/LyNoS

---
license: mit
task_categories:
- image-segmentation
language:
- en
tags:
- medical
pretty_name: AeroPath
size_categories:
- 1B<n<10B
---

<div align="center">
<h1 align="center">🫁 LyNoS 🤗</h1>
<h3 align="center">A multilabel lymph node segmentation dataset from contrast CT</h3>

**LyNoS** was developed by SINTEF Medical Image Analysis to accelerate medical AI research.

</div>

## [Brief intro](https://github.com/raidionics/LyNoS#brief-intro)

This repository contains the LyNoS dataset described in ["_Mediastinal lymph nodes segmentation using 3D convolutional neural network ensembles and anatomical priors guiding_"](https://doi.org/10.1080/21681163.2022.2043778).
The dataset has now also been uploaded to Zenodo and the Hugging Face Hub enabling users to more easily access the data through Python API.

We have also developed a web demo to enable others to easily test the pretrained model presented in the paper. The application was developed using [Gradio](https://www.gradio.app) for the frontend and the segmentation is performed using the [Raidionics](https://raidionics.github.io/) backend.

## [Dataset](https://github.com/raidionics/LyNoS#data) <a href="https://colab.research.google.com/gist/andreped/274bf953771059fd9537877404369bed/lynos-load-dataset-example.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>

### [Accessing dataset](https://github.com/raidionics/LyNoS#accessing-dataset)

The dataset contains 15 CTs with corresponding lymph nodes, azygos, esophagus, and subclavian carotid arteries. The folder structure is described below.

The easiest way to access the data is through Python with Hugging Face's [datasets](https://pypi.org/project/datasets/) package:
```
from datasets import load_dataset

# downloads data from Zenodo through the Hugging Face hub
# - might take several minutes (~5 minutes in CoLab)
dataset = load_dataset("andreped/LyNoS")
print(dataset)

# list paths of all available patients and corresponding features (ct/lymphnodes/azygos/brachiocephalicveins/esophagus/subclaviancarotidarteries)
for d in dataset["test"]:
print(d)
```

A detailed interactive demo on how to load and work with the data can be seen on CoLab. Click the CoLab badge <a href="https://colab.research.google.com/gist/andreped/274bf953771059fd9537877404369bed/lynos-load-dataset-example.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a> to see the notebook or alternatively click [here](https://github.com/raidionics/LyNoS/blob/main/notebooks/lynos-load-dataset-example.ipynb) to see it on GitHub.

### [Dataset structure](https://github.com/raidionics/LyNoS#dataset-structure)

```
└── LyNoS.zip
├── stations_sto.csv
└── LyNoS/
├── Pat1/
│ ├── pat1_data.nii.gz
│ ├── pat1_labels_Azygos.nii.gz
│ ├── pat1_labels_Esophagus.nii.gz
│ ├── pat1_labels_LymphNodes.nii.gz
│ └── pat1_labels_SubCarArt.nii.gz
├── [...]
└── Pat15/
├── pat15_data.nii.gz
├── pat15_labels_Azygos.nii.gz
├── pat15_labels_Esophagus.nii.gz
├── pat15_labels_LymphNodes.nii.gz
└── pat15_labels_SubCarArt.nii.gz
```

### [NIH Dataset Completion](https://github.com/raidionics/LyNoS#nih-dataset-completion)
A larger dataset made of 90 patients featuring enlarged lymph nodes has also been made available by the National Institutes of Health, and is available for download on the official [web-page](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=19726546).
As a supplement to this dataset, lymph nodes segmentation masks have been refined for all patients and stations have been manually assigned to each, available [here](https://drive.google.com/uc?id=1iVCnZc1GHwtx9scyAXdANqz2HdQArTHn).

## [Demo](https://github.com/raidionics/LyNoS#demo) <a target="_blank" href="https://huggingface.co/spaces/andreped/LyNoS"><img src="https://img.shields.io/badge/🤗%20Hugging%20Face-Spaces-yellow.svg"></a>

To access the live demo, click on the `Hugging Face` badge above. Below is a snapshot of the current state of the demo app.

<img width="1400" alt="Screenshot 2023-11-09 at 20 53 29" src="https://github.com/raidionics/LyNoS/assets/29090665/ce661da0-d172-4481-b9b5-8b3e29a9fc1f">

## [Development](https://github.com/raidionics/LyNoS#development)

### [Docker](https://github.com/raidionics/LyNoS#docker)

Alternatively, you can deploy the software locally. Note that this is only relevant for development purposes. Simply dockerize the app and run it:

```
docker build -t LyNoS .
docker run -it -p 7860:7860 LyNoS
```

Then open `http://127.0.0.1:7860` in your favourite internet browser to view the demo.

### [Python](https://github.com/raidionics/LyNoS#python)

It is also possible to run the app locally without Docker. Just setup a virtual environment and run the app.
Note that the current working directory would need to be adjusted based on where `LyNoS` is located on disk.

```
git clone https://github.com/raidionics/LyNoS.git
cd LyNoS/

virtualenv -python3 venv --clear
source venv/bin/activate
pip install -r ./demo/requirements.txt

python demo/app.py --cwd ./
```

## [Citation](https://github.com/raidionics/LyNoS#citation)

If you found the dataset and/or web application relevant in your research, please cite the following reference:
```
@article{bouget2021mediastinal,
author = {David Bouget and André Pedersen and Johanna Vanel and Haakon O. Leira and Thomas Langø},
title = {Mediastinal lymph nodes segmentation using 3D convolutional neural network ensembles and anatomical priors guiding},
journal = {Computer Methods in Biomechanics and Biomedical Engineering: Imaging \& Visualization},
volume = {0},
number = {0},
pages = {1-15},
year = {2022},
publisher = {Taylor & Francis},
doi = {10.1080/21681163.2022.2043778},
URL = {https://doi.org/10.1080/21681163.2022.2043778},
eprint = {https://doi.org/10.1080/21681163.2022.2043778}
}
```

## [License](https://github.com/raidionics/LyNoS#license)

The code in this repository is released under [MIT license](https://github.com/raidionics/LyNoS/blob/main/LICENSE).

---
许可证：MIT协议
任务类别：图像分割
语言：英语
标签：医学
数据集昵称：AeroPath
数据规模：10亿 < 样本量 < 100亿
---

<div align="center">
<h1 align="center">🫁 LyNoS 🤗</h1>
<h3 align="center">基于增强CT的多标签淋巴结分割数据集</h3>

**LyNoS**由SINTEF医学图像分析部门（SINTEF Medical Image Analysis）开发，旨在加速医学人工智能研究的发展。

</div>

## [简介](https://github.com/raidionics/LyNoS#brief-intro)

本仓库包含LyNoS数据集，相关研究成果发表于论文《基于3D卷积神经网络集成与解剖先验引导的纵隔淋巴结分割》（*Mediastinal lymph nodes segmentation using 3D convolutional neural network ensembles and anatomical priors guiding*），DOI: 10.1080/21681163.2022.2043778。目前该数据集已上传至Zenodo和Hugging Face Hub，用户可通过Python接口更便捷地获取数据。

团队还开发了网页演示工具，方便其他研究者快速测试论文中提出的预训练模型。该应用前端基于Gradio框架开发，分割推理后端则依托Raidionics工具库实现。

## [数据集](https://github.com/raidionics/LyNoS#data) <a href="https://colab.research.google.com/gist/andreped/274bf953771059fd9537877404369bed/lynos-load-dataset-example.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="在Colab中打开"/></a>

### [数据集获取](https://github.com/raidionics/LyNoS#accessing-dataset)

该数据集包含15例带标注的CT（计算机断层扫描）影像，标注结构涵盖淋巴结、奇静脉、食管以及锁骨下颈动脉。数据集的文件夹结构如下所述。

最便捷的数据获取方式是通过Python结合Hugging Face的[datasets](https://pypi.org/project/datasets/)工具包：
python
from datasets import load_dataset

# 从Hugging Face Hub通过Zenodo下载数据
# - 在Colab环境中约需5分钟
dataset = load_dataset("andreped/LyNoS")
print(dataset)

# 列出所有患者的路径及对应特征（CT影像/淋巴结标注/奇静脉标注/头臂静脉标注/食管标注/锁骨下颈动脉标注）
for d in dataset["test"]:
print(d)

关于如何加载与使用该数据集的交互式详细演示可在Colab中查看。点击上方的Colab徽章即可打开对应笔记本，或点击[此处](https://github.com/raidionics/LyNoS/blob/main/notebooks/lynos-load-dataset-example.ipynb)跳转至GitHub查看。

### [数据集结构](https://github.com/raidionics/LyNoS#dataset-structure)

└── LyNoS.zip
├── stations_sto.csv
└── LyNoS/
├── Pat1/
│ ├── pat1_data.nii.gz
│ ├── pat1_labels_Azygos.nii.gz
│ ├── pat1_labels_Esophagus.nii.gz
│ ├── pat1_labels_LymphNodes.nii.gz
│ └── pat1_labels_SubCarArt.nii.gz
├── [...]
└── Pat15/
├── pat15_data.nii.gz
├── pat15_labels_Azygos.nii.gz
├── pat15_labels_Esophagus.nii.gz
├── pat15_labels_LymphNodes.nii.gz
└── pat15_labels_SubCarArt.nii.gz

### [美国国立卫生研究院数据集补充](https://github.com/raidionics/LyNoS#nih-dataset-completion)

美国国立卫生研究院（National Institutes of Health, NIH）还发布了一个包含90例淋巴结肿大患者的更大规模数据集，可通过其官方[网页](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=19726546)下载。作为该数据集的补充，本团队已为所有患者的淋巴结分割掩码进行了精细化处理，并手动为每个淋巴结标注了分区信息，相关文件可通过[此处](https://drive.google.com/uc?id=1iVCnZc1GHwtx9scyAXdANqz2HdQArTHn)获取。

## [演示工具](https://github.com/raidionics/LyNoS#demo) <a target="_blank" href="https://huggingface.co/spaces/andreped/LyNoS"><img src="https://img.shields.io/badge/🤗%20Hugging%20Face-Spaces-yellow.svg"></a>

如需使用在线演示工具，请点击上方的Hugging Face徽章。下图为当前演示应用的界面快照。

<img width="1400" alt="Screenshot 2023-11-09 at 20 53 29" src="https://github.com/raidionics/LyNoS/assets/29090665/ce661da0-d172-4481-b9b5-8b3e29a9fc1f">

## [开发部署](https://github.com/raidionics/LyNoS#development)

### [Docker部署](https://github.com/raidionics/LyNoS#docker)

你也可以在本地部署该软件，请注意该部署方式仅适用于开发场景。只需将应用容器化并运行即可：
bash
docker build -t LyNoS .
docker run -it -p 7860:7860 LyNoS

随后在你偏好的浏览器中打开`http://127.0.0.1:7860`即可访问演示应用。

### [Python本地部署](https://github.com/raidionics/LyNoS#python)

无需Docker亦可在本地运行该应用，只需创建虚拟环境并运行应用即可。请注意需根据`LyNoS`在本地磁盘中的路径调整当前工作目录。

bash
git clone https://github.com/raidionics/LyNoS.git
cd LyNoS/

virtualenv -python3 venv --clear
source venv/bin/activate
pip install -r ./demo/requirements.txt

python demo/app.py --cwd ./

随后在你偏好的浏览器中打开`http://127.0.0.1:7860`即可访问演示应用。

## [引用方式](https://github.com/raidionics/LyNoS#citation)

若您的研究中使用了本数据集或网页演示工具，请引用以下文献：
bibtex
@article{bouget2021mediastinal,
author = {David Bouget and André Pedersen and Johanna Vanel and Haakon O. Leira and Thomas Langø},
title = {Mediastinal lymph nodes segmentation using 3D convolutional neural network ensembles and anatomical priors guiding},
journal = {Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization},
volume = {0},
number = {0},
pages = {1-15},
year = {2022},
publisher = {Taylor & Francis},
doi = {10.1080/21681163.2022.2043778},
URL = {https://doi.org/10.1080/21681163.2022.2043778},
eprint = {https://doi.org/10.1080/21681163.2022.2043778}
}

## [许可证](https://github.com/raidionics/LyNoS#license)

本仓库中的代码采用[MIT协议](https://github.com/raidionics/LyNoS/blob/main/LICENSE)开源发布。

原始信息汇总

数据集概述

基本信息

• 许可证: MIT
• 任务类别: 图像分割
• 语言: 英语
• 标签: 医疗
• 易读名称: AeroPath
• 数据集大小: 1B<n<10B

数据集描述

LyNoS 是由 SINTEF Medical Image Analysis 开发的用于加速医学AI研究的多标签淋巴结分割数据集，基于对比CT图像。

数据集访问

数据集包含15个CT图像及其对应的淋巴结、奇静脉、食道和锁骨下颈动脉。可以通过Python使用Hugging Face的datasets包来访问数据集： python from datasets import load_dataset

dataset = load_dataset("andreped/LyNoS") print(dataset)

for d in dataset["test"]: print(d)

数据集结构

数据集结构如下：

└── LyNoS.zip ├── stations_sto.csv └── LyNoS/ ├── Pat1/ │ ├── pat1_data.nii.gz │ ├── pat1_labels_Azygos.nii.gz │ ├── pat1_labels_Esophagus.nii.gz │ ├── pat1_labels_LymphNodes.nii.gz │ └── pat1_labels_SubCarArt.nii.gz ├── [...] └── Pat15/ ├── pat15_data.nii.gz ├── pat15_labels_Azygos.nii.gz ├── pat15_labels_Esophagus.nii.gz ├── pat15_labels_LymphNodes.nii.gz └── pat15_labels_SubCarArt.nii.gz

引用

如果数据集和/或网络应用对你的研究有帮助，请引用以下参考文献：

@article{bouget2021mediastinal, author = {David Bouget and André Pedersen and Johanna Vanel and Haakon O. Leira and Thomas Langø}, title = {Mediastinal lymph nodes segmentation using 3D convolutional neural network ensembles and anatomical priors guiding}, journal = {Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization}, volume = {0}, number = {0}, pages = {1-15}, year = {2022}, publisher = {Taylor & Francis}, doi = {10.1080/21681163.2022.2043778}, URL = {https://doi.org/10.1080/21681163.2022.2043778}, eprint = {https://doi.org/10.1080/21681163.2022.2043778} }

许可证

本仓库中的代码在MIT许可证下发布。

搜集汇总

背景与挑战

背景概述

LyNoS是一个用于医学AI研究的淋巴节点多标签分割数据集，包含15个CT扫描图像及其对应的淋巴节点、奇静脉、食管和锁骨下颈动脉标签数据。数据集支持通过Hugging Face的datasets包便捷访问，并附有相关研究论文和在线演示应用。