- Jupyter Notebook 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| Codes | ||
| Models/Svetlana | ||
| Setup | ||
| How to re-train the classifier.md | ||
| LICENSE | ||
| MyoFuse : How to Use.md | ||
| README.md | ||
MyoFuse
MyoFuse is a fully AI-based workflow to quantify the fusion index in skeletal muscle cell 2D culture in vitro. The workflow is currently at version v1.0.0.
In this repository, you will find all the necessary models, code, and instructions for use. The images used to train the classifier are available on Zenodo.
How to Use MyoFuse
Detailed instructions to use the workflow are provided in the MyoFuse : How to Use.md guide. Instructions for retraining the classifier are also available. Here's a quick overview of the setup process:
-
Setup Environment:
In theSetupfolder, you will find `requirements' files that list the necessary dependencies, depending on your OS. Use it to set up your Python environment.
This will also install napari and the svetlana plugin if you need to train your own model or retrain the classifier. -
Download Models: The classifier model,
MyoFuse.pth, is available in theModelssection of this repository. Download it and place it in the appropriate folder as described in the instructions. If you are using the Svetlana plugin for classification, add theSvetlanafolder to your project directory, as explained in the instructions. The segmentation model for Cellpose is included in the first release of this repository. -
Process Your Images: Use the Python notebooks provided in the
Codessection to process your images and quantify the fusion index. Follow the step-by-step guideMyoFuse: How to Useto ensure accurate results. -
Test Images: If you want to make sure that your configuration is working, you can download test images from the
Test Imagesrelease to run theMyoFuse.ipynbcode. A large image is also provided to test potential RAM and GPU memory issues.
Licence
This project is licensed under the MIT License.
Citation
If you use MyoFuse, please cite ...
Contact
If you have any questions or suggestions, feel free to reach out:
Email: lair.b1jamin@gmail.com