itsasimiqbal/sebre

Developing a Brain Atlas through Deep Learning

SeBRe 2.0 PyTorch support (2023 - New)!

SeBRe (Segmenting Brain Regions) is a high-throughput Mask R-CNN [1]-based toolbox to generate brain atlas through deep learning by instance segmentation of complex brain regions.

Nature Machine Intelligence article: Asim Iqbal, et al. Developing a brain atlas through deep learning. Nature Machine Intelligence

arXiv preprint: SeBRe

The Colab version of SeBRe is available here to train and test on your custom dataset:

SeBRe 1.0 TensorFlow support!

Download:

Clone the SeBRe repository by running the following command in your terminal window:

git clone https://github.com/itsasimiqbal/SeBRe.git

Additionally, download the (.zip) files by visiting the following link:

https://drive.google.com/open?id=1g_894zM_mSJGfudzzfnw-ZpN-rQ4qQdz

You would need DATASETSubmit.zip, myDATASET.zip and SeBRe_FINAL_WEIGHTS.h5.zip to run the code in your machine. If you'd like to use the mouse and human brain atlas datasets used in our paper then also download and unzip SeBRe_Datasets.zip. Once you've downloaded your files then unzip them in your folder by running the following commands:

unzip DATASETSubmit.zip
unzip myDATASET.zip
unzip SeBRe_FINAL_WEIGHTS.h5.zip

Following are the Python/library versions on which the SeBRe 1.0 code is tested to work fine:

• Python (3.5.2)
• Tensorflow (1.6.0)
• Keras (2.1.6)
• skimage (0.13.0)
• Numpy (1.13.3)
• Scipy (1.2.2)

SeBRe demo:

The block diagram of our system is demonstrated below:

Run the notebook SeBRe_FINAL.ipynb to reproduce the results in SeBRe's paper. Make sure to install the necessary libraries in your machine before running the code. A step-by-step explanation of feature processing in SeBRe is provided in SeBRe_feature_processing.ipynb notebook.

To train SeBRe on your (custom) dataset:

1. Collect images of brain regions (or sub-regions) and place each of them in a separate folder with the following naming convention e.g. section_img_0.jpg, section_img_1.jpg, ...

2. Draw ground-truth (binary) masks on Regions of Interest (ROIs) e.g. cortex, hippocampus, etc. using an annotation software such as Napari [2] and assign a unique color code to each ROI. In the figure below, mouse brain section is shown before (a) and after (d) annotation by human expert. A zoomed-in examples of cortex (b) and hindbrain (c) are shown to precisely match the boundaries of masks with the corresponding ROIs.

3. Run the notebook custom_dataset_create.ipynb to generate the binary (black and white) masks to train SeBRe deep neural network (DNN). The notebook will generate the file names in SeBRe-readable format, place them in the brain region corresponding folders e.g. section_masks_0/section_masks_0_m_1.png, sections_masks_0_m_2.png, ...

4. Modify and run the notebook SeBRe_training.ipynb to train the SeBRe DNN on your (custom) dataset.

5. Modify and run the notebook SeBRe_FINAL.ipynb to test the SeBRe DNN on your (custom) dataset.

References:

[1] https://arxiv.org/abs/1703.06870

[2] https://napari.org/

Cite:

If you use any part of this code for your work, please cite the following:

Asim Iqbal, Romesa Khan, and Theofanis Karayannis. "Developing a brain atlas through deep learning." 
Nature Machine Intelligence 1.6 (2019): 277-287.

sebre's Issues

Coronal Segmentation

Hello All,
I noticed in the paper published in Nature that you guys performed testing on coronal scans.
Could I get in contact with someone to get the weights from training on the Allen Atlas?
Thanks!

label:enhancement Add Elastix, and Ndreg parameters used in the paper

It would be interesting to know what Elastix and Ndreg (if non default) parameters were used in the paper.

Cannot download data from google drive

I tried downloading data and gets 3rd party cookies error, which is already allowed in settings.

Compatibility Issues

Hi, I am also trying to reproduce the results of your code and am having multiple compatibility issues. Would you be able to share the version of python you used and the versions of all the packages you used?

Keras package compatibility

Hi, I am trying to reproduce the results of your code but ran into some difficulty. May I know which tensorflow and keras packages you used in your code. I tried using keras 2.0.8 which comes with tensorflow 1.10.0 but encountered an error in the Supplementary Figure 4(b) section of the code in SeBRe_feature_processing.ipynb. I got an AssertionError when defining rpn.

~\SeBRe\model.py in run_graph(self, images, outputs)
   2448         outputs = OrderedDict(outputs)
   2449         for o in outputs.values():
-> 2450             assert o is not None

I'm very new to coding and python. Any help is appreciated. Thanks.