StARQ: brainStem Automated Registration and Quantification

StARQ is an open-source software for automated segmentation and quantification of the brainstem structures in imaging data. StARQ is written in Python and built with PyTorch.

Features

- Automated segmentation of the brainstem structures in 2D mouse brain images
- Quantification of cross-sectional regions in brainstem and signal quantification
- Interactive visualization of the processed results with save option

Read our pre-print on bioRxiv

Here's the link to our pre-print article:

Usage

To use StARQ, you will need to have a set of mouse brainstem images in .tif format (with registration and signal channels). Once your data is prepared, you can run StARQ by executing the following scripts:

Finetune on your custom (mouse braistem imaging) dataset

Run the Google Colab to test it on a sample mouse brainstem imaging data as well as fine-tune on your custom data. The input_directory should be changed with your data, and output_directory should be where you want the results to be saved.

Run on your custom (mouse braistem imaging) dataset

Run this Google Colab to load your custom image data (.tif files with registration and signal channels) from Google Drive and pass it to the StARQ deep learning framework, and save your signal quantification results.

Contributing

If you are interested in contributing to StARQ, please fork the repository and submit a pull request. We welcome contributions of bug fixes, new features, and improvements to the documentation.

License

StARQ is released under the MIT License.

Citation

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

@article{kaiser2023domain,
  title={Domain-Invariant Brainstem Nuclei Segmentation and Signal Quantification},
  author={Kaiser, Julia and Luong, Dana and Sung, Eunseo and Iqbal, Asim and Sahni, Vibhu},
  journal={bioRxiv},
  pages={2023--11},
  year={2023},
  publisher={Cold Spring Harbor Laboratory}
}