Model Overview

This repository contains the RSS-Net framework for the task of rodent MRI skull stripping

Installing Dependencies

Our framework is developed based on Python 3.9, PyTorch 2.0.0, MONAI, and nnUNet.

Dependencies can be installed using:

conda create -n rss python=3.9

conda activate rss

# CUDA 11.7
conda install pytorch==2.0.0 torchvision==0.15.0 torchaudio==2.0.0 pytorch-cuda=11.7 -c pytorch -c nvidia
# CUDA 11.8
conda install pytorch==2.0.0 torchvision==0.15.0 torchaudio==2.0.0 pytorch-cuda=11.8 -c pytorch -c nvidia
# CPU Only
conda install pytorch==2.0.0 torchvision==0.15.0 torchaudio==2.0.0 cpuonly -c pytorch

pip install -r requirements.txt

python setup.py install

Model

Our network architecture is illustrated in the following diagram:

Data Description

We trained the model using MRI data from rodent brains collected from 89 centers. You can download the pre-trained model files and the annotated brain masks Google Drive or MEGA.

You can find the download links for these datasets in our article (in submission)

Dataset		MultiRat	StdRat	C Rat	TRat	C Mouse	NeAt
Participant		646	209	132	24	16	10
Image		751	209	132	24	16	10
Sex	Male	505	116	132		16	10
	Female	141	93		24
Strain	Wistar	256	189	16	24
	SD	220		94
	LE	80	10	22
	F344	60	10
	LH	30
	WTC					16
	IC						10
Age(months)	0-2	164	88
	2-4	308	57				12-14 weeks
	4-6	16	2
	6-12	22	2
	12-20	28
	-	108	60	adult	adult	adult
Weight(grams)	100-200	71	20
	200-250	78	71
	250-300	97	56
	300-350	173	40
	350-400	81	13
	400-700	94	9
	-	52		132		16	25-30 grams
Vendor	Bruker	568	180	132	24	16	10
	MS	10
	SD	68
	A/B		10
	Varian		10
	Clinscan		9
Field Strength(T)	3	10
	4.7	40	10
	7	309	79
	9.4	229	100	132	24	16	10
	11.7	30
	14		10
	14.1	28
	17.2		10

Usage

After the installation is complete, you can use the following command to run the framework:

 RS2_predict -i 'path/to/input' -o 'path/to/output' -m 'path/to/pretrained_model.pt'

We recommend using absolute paths to avoid unnecessary problem.

To meet the required naming format, ensure that your input files are organized as follows:

|input_folder_path
|---- <file-name1>_0000.nii.gz
|---- <file-name2>_0000.nii.gz

Additional Reminders

1. In December 2024, the library "acvl_utils" was updated (version 0.2.2), which led to errors when used with Python 3.9(this issue). Please install an earlier version of "acvl_utils", such as version 0.2.1.
2. If you encounter the following problem when using it: FileNotFoundError: [Errno 2] No such file or directory: 'your_conda_path/envs/rss/lib/python3.9/site-packages/RS2-1.0-py3.9.egg/RS2/jsons/dataset.json' Please copy all the files from "Rodent-Skull-Stripping/RS2/jsons" to "your_conda_path/envs/rss/lib/python3.9/site-packages/RS2-1.0-py3.9.egg/RS2/jsons". This should solve the problem.

Comand Parameters

Here are the parameters for the RS2_predict script in Markdown format:

• -i: Input folder. Ensure that the files are in the .nii.gz format and end with _0000.nii.gz.

• -o: Output folder. If it does not exist, it will be created. Predicted segmentations will have the same name as their source images.

• -m: Path of the pretrained model you want to use. Default: RS2_pretrained_model.pt.

• -device: Use this to set the device the inference should run with. Available options are 'cuda' (GPU), 'cpu' (CPU), and 'mps' (Apple M1/M2). Do NOT use this to set which GPU ID! Use CUDA_VISIBLE_DEVICES=X RS2_predict [...] instead.

• -step_size: Step size for sliding window prediction. The larger it is, the faster but less accurate the prediction. Default: 0.5. Cannot be larger than 1. We recommend the default.

• --disable_tta: Set this flag to disable test time data augmentation in the form of mirroring. Faster, but less accurate inference. Not recommended.

• --verbose: Set this if you like being talked to. You will have to be a good listener/reader.

• --save_probabilities: Set this to export predicted class "probabilities". Required if you want to ensemble multiple configurations.

• --continue_prediction: Continue an aborted previous prediction (will not overwrite existing files).

• -npp: Number of processes used for preprocessing. More is not always better. Beware of out-of-RAM issues. Default: 3.

• -nps: Number of processes used for segmentation export. More is not always better. Beware of out-of-RAM issues. Default: 3.

Citation

If you find this repository useful, please consider citing our paper:

@article{lin2024rs2,
  title={RS2-Net: An end-to-end deep learning framework for rodent skull stripping in multi-center brain MRI},
  author={Lin, Yongkang and Ding, Yanhui and Chang, Shulei and Ge, Xinting and Sui, Xiaodan and Jiang, Yanyun},
  journal={NeuroImage},
  volume={298},
  pages={120769},
  year={2024},
  publisher={Elsevier}
}