You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CITATION.cff
+5-3Lines changed: 5 additions & 3 deletions
Original file line number
Diff line number
Diff line change
@@ -13,7 +13,9 @@ authors:
13
13
- family-names: Caan
14
14
given-names: Matthan
15
15
orcid: https://orcid.org/0000-0002-5162-8880
16
-
title: "Advanced Toolbox for Multitask Medical Imaging Consistency (ATOMMIC)"
16
+
title: "Atommic: An Advanced Toolbox for Multitask Medical Imaging Consistency to Facilitate Artificial Intelligence Applications from Acquisition to Analysis in Magnetic Resonance Imaging"
The [Advanced Toolbox for Multitask Medical Imaging Consistency (ATOMMIC)](https://github.com/wdika/atommic) is a toolbox for applying AI methods for **accelerated MRI reconstruction (REC)**, **MRI segmentation (SEG)**, **quantitative MR imaging (qMRI)**, as well as **multitask learning (MTL)**, i.e., performing multiple tasks simultaneously, such as reconstruction and segmentation. Each task is implemented in a separate collection consisting of data loaders, transformations, models, metrics, and losses. **ATOMMIC** is designed to be modular and extensible on new tasks, models, and datasets. **ATOMMIC** uses [PyTorch Lightning](https://www.pytorchlightning.ai/) for feasible high-performance multi-GPU/multi-node mixed-precision training.
The schematic overview of **ATOMMIC** showcases the main components of the toolbox. First, we need an [MRI Dataset](README.md#mri-datasets) (e.g., **CC359**). Next, we need to define the high-level parameters, such as the [task and the model](https://atommic.readthedocs.io/en/latest/mri/collections.html), the [undersampling](https://atommic.readthedocs.io/en/latest/mri//undersampling.html), the [transforms](https://atommic.readthedocs.io/en/latest/mri//transforms.html), the [optimizer](https://atommic.readthedocs.io/en/latest/core/core.html#optimization), the [scheduler](https://atommic.readthedocs.io/en/latest/core/core.html#learning-rate-schedulers), the [loss](https://atommic.readthedocs.io/en/latest/mri/losses.html), the [trainer parameters](https://atommic.readthedocs.io/en/latest/core/core.html#training), and the [experiment manager](https://atommic.readthedocs.io/en/latest/core/exp_manager.html). All these parameters are defined in a `.yaml` file using [Hydra](https://hydra.cc/) and [OmegaConf](https://omegaconf.readthedocs.io/).
28
33
29
34
The trained model is an `.atommic`[module](https://atommic.readthedocs.io/en/latest/core/export.html), exported with [ONNX](https://onnx.ai/) and [TorchScript](https://pytorch.org/docs/stable/jit.html) support, which can be used for inference. The `.atommic` module can also be uploaded on [HuggingFace](https://huggingface.co/). Pretrained models are available on our [HF](https://huggingface.co/wdika) account and can be downloaded and used for inference.
30
35
36
+
## 🛠️ Installation
37
+
38
+
**ATOMMIC** is best to be installed in a Conda environment.
39
+
40
+
### 🐍 Conda
41
+
```
42
+
conda create -n atommic python=3.10
43
+
conda activate atommic
44
+
```
45
+
46
+
### 📦 Pip
47
+
Use this installation mode if you want the latest released version.
48
+
49
+
```bash
50
+
pip install atommic
51
+
```
52
+
53
+
### From source
54
+
55
+
Use this installation mode if you are contributing to atommic.
56
+
57
+
```bash
58
+
git clone https://github.com/wdika/atommic
59
+
cd atommic
60
+
bash ./reinstall.sh
61
+
```
62
+
63
+
### 🐳 Docker containers
64
+
65
+
An atommic container is available at dockerhub, you can pull it with:
docker run --gpus all -it --rm -v /home/user/configs:/config atommic:latest atommic run -c /config/config.yaml
78
+
```
79
+
where ```/config/config.yaml``` is the path to your local configuration file.
80
+
81
+
Or you can run it interactively with:
82
+
```bash
83
+
docker run --gpus all -it --rm -p 8888:8888 atommic:latest /bin/bash -c "./start-jupyter.sh"
84
+
```
85
+
31
86
## 🚀 Quick Start Guide
32
87
33
-
The best way to get started with ATOMMIC is to start with one of the [tutorials](https://atommic.readthedocs.io/en/latest/starthere/tutorials.html):
88
+
The best way to get started with ATOMMIC is with one of the [tutorials](https://atommic.readthedocs.io/en/latest/starthere/tutorials.html):
34
89
35
90
-[ATOMMIC Primer](https://github.com/wdika/atommic/blob/main/tutorials/00_ATOMMIC_Primer.ipynb) - demonstrates how to use ATOMMIC.
36
91
-[ATOMMIC MRI transforms](https://github.com/wdika/atommic/blob/main/tutorials/01_ATOMMIC_MRI_transforms.ipynb) - demonstrates how to use ATOMMIC to undersample MRI data.
@@ -39,12 +94,13 @@ The best way to get started with ATOMMIC is to start with one of the [tutorials]
39
94
40
95
You can also check the [projects](https://github.com/wdika/atommic/tree/main/projects) page to see how to use ATOMMIC for specific tasks and public datasets.
41
96
97
+
Pre-trained models are available on HuggingFace [🤗](https://huggingface.co/wdika).
98
+
42
99
### **ATOMMIC paper is fully reproducible. Please check [here](https://github.com/wdika/atommic/tree/main/projects/ATOMMIC_paper/README.md) for more information.**
43
100
44
101
## 🤖 Training & Testing
45
102
46
-
Training and testing models in **ATOMMIC** is intuitive and easy. You just need to properly configure the `.yaml`
47
-
file and just run the following command:
103
+
Training and testing models in **ATOMMIC** is intuitive and easy. You just need to properly configure a `.yaml` file and run the following command:
48
104
49
105
```bash
50
106
atommic run -c path-to-config-file
@@ -102,48 +158,6 @@ You can also check the [projects](https://github.com/wdika/atommic/tree/main/pro
102
158
-[SKM-TEA](https://github.com/wdika/atommic/tree/main/projects/MTL/rs/SKMTEA): Supports the `(REC)`, `(SEG)`, and `(MTL)` tasks.
103
159
-[Stanford Knees](https://github.com/wdika/atommic/tree/main/projects/REC/StanfordKnees2019): Supports the `(REC)` task.
104
160
105
-
## 🛠️ Installation
106
-
107
-
**ATOMMIC** is best to be installed in a Conda environment.
108
-
109
-
### 🐍 Conda
110
-
```
111
-
conda create -n atommic python=3.10
112
-
conda activate atommic
113
-
```
114
-
115
-
### 📦 Pip
116
-
Use this installation mode if you want the latest released version.
117
-
118
-
```bash
119
-
pip install atommic
120
-
```
121
-
122
-
### From source
123
-
124
-
Use this installation mode if you are contributing to atommic.
125
-
126
-
```bash
127
-
git clone https://github.com/wdika/atommic
128
-
cd atommic
129
-
bash ./reinstall.sh
130
-
```
131
-
132
-
### 🐳 Docker containers
133
-
To build an atommic container with Dockerfile from a branch, please run
As [NeMo](https://github.com/NVIDIA/NeMo) suggests, if you choose to work with the `main` branch, use NVIDIA's PyTorch container version [21.05-py3](https://ngc.nvidia.com/containers/nvidia:pytorch/tags), then install from GitHub.
140
-
141
-
```bash
142
-
docker run --gpus all -it --rm -v <atommic_github_folder>:/ATOMMIC --shm-size=8g \
title = {ATOMMIC: Advanced Toolbox for Multitask Medical Imaging Consistency},
166
-
year = {2023},
167
-
url = {https://github.com/wdika/atommic},
168
-
}
176
+
@article{Karkalousos_2024,
177
+
title={Atommic: An Advanced Toolbox for Multitask Medical Imaging Consistency to Facilitate Artificial Intelligence Applications from Acquisition to Analysis in Magnetic Resonance Imaging},
178
+
url={http://dx.doi.org/10.2139/ssrn.4801289},
179
+
DOI={10.2139/ssrn.4801289},
180
+
publisher={Elsevier BV},
181
+
author={Karkalousos, Dimitrios and Išgum, Ivana and Marquering, Henk and Caan, Matthan W.A.},
182
+
year={2024}}
169
183
```
170
184
171
185
## 🔗 References
172
186
173
-
The following papers have used ATOMMIC:
187
+
ATOMMIC has been used or is referenced in the following papers:
188
+
189
+
1. Karkalousos, Dimitrios and Išgum, Ivana and Marquering, Henk and Caan, Matthan W.A., Atommic: An Advanced Toolbox for Multitask Medical Imaging Consistency to Facilitate Artificial Intelligence Applications from Acquisition to Analysis in Magnetic Resonance Imaging. Available at SSRN: https://ssrn.com/abstract=4801289 or http://dx.doi.org/10.2139/ssrn.4801289
190
+
191
+
2. Karkalousos, D., Išgum, I., Marquering, H. A., & Caan, M. W. A. (2024). ATOMMIC: An Advanced Toolbox for Multitask Medical Imaging Consistency to facilitate Artificial Intelligence applications from acquisition to analysis in Magnetic Resonance Imaging. https://doi.org/10.2139/ssrn.4801289
192
+
193
+
3. Karkalousos, D., Isgum, I., Marquering, H., & Caan, M. W. A. (2024, April 27). The Advanced Toolbox for Multitask Medical Imaging Consistency (ATOMMIC): A Deep Learning framework to facilitate Magnetic Resonance Imaging. Medical Imaging with Deep Learning. https://openreview.net/forum?id=HxTZr9yA0N
174
194
175
-
1. Karkalousos, D., Isgum, I., Marquering, H. & Caan, M.W.A.. (2024). MultiTask Learning for accelerated-MRI Reconstruction and Segmentation of Brain Lesions in Multiple Sclerosis. <i>Medical Imaging with Deep Learning</i>, in <i>Proceedings of Machine Learning Research</i> 227:991-1005 Available from https://proceedings.mlr.press/v227/karkalousos24a.html.
195
+
4. Karkalousos, D., Isgum, I., Marquering, H. & Caan, M.W.A.. (2024). MultiTask Learning for accelerated-MRI Reconstruction and Segmentation of Brain Lesions in Multiple Sclerosis. <i>Medical Imaging with Deep Learning</i>, in <i>Proceedings of Machine Learning Research</i> 227:991-1005 Available from https://proceedings.mlr.press/v227/karkalousos24a.html.
176
196
177
-
2. Zhang, C., Karkalousos, D., Bazin, P. L., Coolen, B. F., Vrenken, H., Sonke, J. J., Forstmann, B. U., Poot, D. H. J., & Caan, M. W. A. (2022). A unified model for reconstruction and R2* mapping of accelerated 7T data using the quantitative recurrent inference machine. NeuroImage, 264. [DOI](https://doi.org/10.1016/j.neuroimage.2022.119680)
197
+
5. Zhang, C., Karkalousos, D., Bazin, P. L., Coolen, B. F., Vrenken, H., Sonke, J. J., Forstmann, B. U., Poot, D. H. J., & Caan, M. W. A. (2022). A unified model for reconstruction and R2* mapping of accelerated 7T data using the quantitative recurrent inference machine. NeuroImage, 264. [DOI](https://doi.org/10.1016/j.neuroimage.2022.119680)
178
198
179
-
3. Karkalousos, D., Noteboom, S., Hulst, H. E., Vos, F. M., & Caan, M. W. A. (2022). Assessment of data consistency through cascades of independently recurrent inference machines for fast and robust accelerated MRI reconstruction. Physics in Medicine & Biology. [DOI](https://doi.org/10.1088/1361-6560/AC6CC2)
199
+
6. Karkalousos, D., Noteboom, S., Hulst, H. E., Vos, F. M., & Caan, M. W. A. (2022). Assessment of data consistency through cascades of independently recurrent inference machines for fast and robust accelerated MRI reconstruction. Physics in Medicine & Biology. [DOI](https://doi.org/10.1088/1361-6560/AC6CC2)
180
200
181
201
## 📧 Contact
182
202
183
203
For any questions, please contact Dimitris Karkalousos @ [d.karkalousos@amsterdamumc.nl](mailto:d.karkalousos@amsterdamumc.nl).
184
204
185
205
## ⚠️🙏 Disclaimer & Acknowledgements
186
206
187
-
> **Note:** ATOMMIC is built on top of [NeMo](https://github.com/NVIDIA/NeMo). NeMo is under Apache 2.0 license, so we are allowed to use it. We also assume that it is allowed to use the NeMo documentation, as long as we cite it and we always refer to the baselines everywhere and in the code and docs. ATOMMIC also includes implementations of reconstruction methods from [fastMRI](https://github.com/facebookresearch/fastMRI) and [DIRECT](https://github.com/NKI-AI/direct), and segmentation methods from [MONAI](https://github.com/Project-MONAI/MONAI), as well as other codebases which are always cited on the corresponding files. All methods in ATOMMIC are reimplemented and not called from the original libraries, allowing for full reproducibility, support, and easy extension. ATOMMIC is an open-source project under the Apache 2.0 license.
207
+
> **Note:** ATOMMIC is built on top of [NeMo](https://github.com/NVIDIA/NeMo). NeMo is under Apache 2.0 license, so we are allowed to use it. We also assume that we can use the NeMo documentation basis as long as we cite it and always refer to the baselines everywhere in the code and docs. ATOMMIC also includes implementations of reconstruction methods from [fastMRI](https://github.com/facebookresearch/fastMRI) and [DIRECT](https://github.com/NKI-AI/direct), and segmentation methods from [MONAI](https://github.com/Project-MONAI/MONAI), as well as other codebases which are always cited on the corresponding files. All methods in ATOMMIC are reimplemented and not called from the original libraries, allowing for full reproducibility, support, and easy extension. ATOMMIC is an open-source project under the Apache 2.0 license.
0 commit comments