Chest-xray-landmark-dataset

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

Chest-xray-landmark-dataset

★ 24 Open GitHub ↗

No description available.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Chest x-ray Landmark Segmentation Dataset.

This git repo contains 911 landmark annotations for chest x-ray images from JSRT, Shenzhen, Montgomery and Padchest datasets.\
We are grateful to Padchest's team for providing people an easy way to download our subset of images, under "Sample 2".

The landmark annotations were obtained with an ensemble of HybridGNet 2-IGSC models, using dense segmentation masks as input.

![Example](https://raw.githubusercontent.com/ngaggion/Chest-xray-landmark-dataset/main/figs/landmarks.png)

Pre-processing scripts are provided for every dataset, please download them from their official websites.\
Links provided in the corresponding .ipynb file.

Lung annotations are available for images from all datasets. \
Heart annotations are only available for JSRT and Padchest images.

Annotations:\
RL : Right Lung (44 landmarks) \
LL : Left Lung (50 landmarks) \
H : Heart (26 landmarks)

![workflow](https://raw.githubusercontent.com/ngaggion/Chest-xray-landmark-dataset/main/figs/info.png)

## New study on combining heterogeneous landmark annotations using this dataset

Please check our latest paper "Multi-center anatomical segmentation with heterogeneous labels via landmark-based models".

Arxiv link: https://arxiv.org/abs/2211.07395

Source code: https://github.com/ngaggion/MultiCenterSegRepo

## Baseline results

Results were taken under a 20% balanced test holdout set. For more details check on https://github.com/ngaggion/Chest-x-ray-Baselines

Results for both lungs and heart segmentation, using the JSRT + Padchest subset:

|      **Model** |        **MSE** | **Dice Lungs** | **Dice Heart** |  **HD Lungs** |  **HD Heart** |
|---------------:|---------------:|---------------:|---------------:|--------------:|--------------:|
|            PCA |     359.696809 |       0.938072 |       0.909292 |     50.469137 |     43.823793 |
|             FC |     359.423800 |       0.938108 |       0.907268 |     50.506582 |     44.751405 |
| **HybridGNet** | **220.685126** |   **0.965563** |   **0.933040** | **37.550593** | **35.982695** |

Results for lung segmentation only, using the full dataset:

|      **Model** |        **MSE** | **Dice Lungs** |  **HD Lungs** |
|---------------:|---------------:|---------------:|--------------:|
|            PCA |     216.482534 |       0.948304 |     40.945048 |
|             FC |     232.796428 |       0.945217 |     42.013886 |
| **HybridGNet** | **164.904743** |   **0.966357** | **34.492442** |

## Citation

If you are using the annotations, please cite our works:

N. Gaggion, L. Mansilla, C. Mosquera, D. H. Milone and E. Ferrante, "Improving anatomical plausibility in medical image segmentation via hybrid graph neural networks: applications to chest x-ray analysis," in IEEE Transactions on Medical Imaging, doi: 10.1109/TMI.2022.3224660.

https://doi.org/10.1109%2Ftmi.2022.3224660

```
@article{Gaggion_2022,
	doi = {10.1109/tmi.2022.3224660},
	url = {https://doi.org/10.1109%2Ftmi.2022.3224660},
	year = 2022,
	publisher = {Institute of Electrical and Electronics Engineers ({IEEE})},
	author = {Nicolas Gaggion and Lucas Mansilla and Candelaria Mosquera and Diego H. Milone and Enzo Ferrante},
	title = {Improving anatomical plausibility in medical image segmentation via hybrid graph neural networks: applications to chest x-ray analysis},
	journal = {{IEEE} Transactions on Medical Imaging}
}
```

Gaggion, N., Mansilla, L., Milone, D. H., & Ferrante, E. (2021, September). Hybrid graph convolutional neural networks for landmark-based anatomical segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 600-610). Springer, Cham.

https://link.springer.com/chapter/10.1007%2F978-3-030-87193-2_57

```
@InProceedings{10.1007/978-3-030-87193-2_57,
author="Gaggion, Nicol{\'a}s
and Mansilla, Lucas
and Milone, Diego H.
and Ferrante, Enzo",
title="Hybrid Graph Convolutional Neural Networks for Landmark-Based Anatomical Segmentation",
booktitle="Medical Image Computing and Computer Assisted Intervention -- MICCAI 2021",
year="2021",
publisher="Springer International Publishing",
address="Cham",
pages="600--610",
isbn="978-3-030-87193-2"
}
```

Chest-xray-landmark-dataset

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README