This repository packages AxonDeepSeg as part of a JupyterLab Service on o²S²PARC. It also adds support for pytorch+GPU/CUDA.
AxonDeepSeg is an open-source software using deep learning and aiming at automatically segmenting axons and myelin sheaths from microscopy images. It performs 3-class semantic segmentation using a convolutional neural network. AxonDeepSeg was developed at NeuroPoly Lab, Polytechnique Montreal, University of Montreal, Canada. See GitHub repository and Documentation for more information.
The base of the Service was built with the cookiecutter-osparc-ui-module, using supervisord, xtigervnc, novnc and openbox
If you use this work in your research, please cite:
Zaimi, A., Wabartha, M., Herman, V., Antonsanti, P.-L., Perone, C. S., & Cohen-Adad, J. (2018). AxonDeepSeg: automatic axon and myelin segmentation from microscopy data using convolutional neural networks. Scientific Reports, 8(1), 3816. Link to the paper.
Build the module:
$ make buildTo run locally at and visit http://127.0.0.1:8888
make run-localTo publish in local throw-away registry:
make publish-localService version is updated with make version-*
A template ci config file is created in .github/workflows/check-image.yml, it checks that the image builds. When the workflow runs successfully for a new version (on the main branch), this is automatically detected and published on the internal registry (see also "Deployment on o²S²PARC" in this README)
The required CI is already packaged. To build and push to the internal registry you must add it to the oSparc/docker-publisher-osparc-services repository.
Run locally and visit http://127.0.0.1:8888:
make run-localOr publish it in a local o²S²PARC deploy:
make publish-localPerform a test segmentation as shown in this video tutorial. You can use as input the image from this repository (from validation/inputs/input_1) or others provided by AxonDeepSeg (e.g. in this repository). The input folder will be mounted in tmp/inputs
First version