myovision

Description

This is the main repository for the Consulting Project: "Quantification of Myogenic Differentiation using Deep Learning", developed at the Ludwig Maximilian University of Munich in partnership with Musculoskeletal Center at the LMU University Hospital.

Supervisors and Project Partner

• Supervisors: Prof. Dr. David Rügamer, Dr. Andreas Bender and Tobias Weber

• Project Partner: Dr. rer. nat. Maximilian Saller

Project Structure

1. Main Repository (this repository) Contains notebooks for presentig main results of the project. Notebooks are located in the notebooks directory:

   • 01_predict_performance.ipynb: Prediction and Performance Computation.
   • 02_exploratory_data_analysis.ipynb: Exploratory Data Analysis
   • 03_classical_cv.ipynb: Classical Computer Vision Techniques

2. Data Collection Contains scripts for collecting and preprocessing training data. For data annotation process we employed a specificly designed tool which is not public for now but we provide a visualization of the data annotation process:

   If you are interested in the Data Annotation Tool, please contact us.
   

   

3. Model Training / Inference Contains the modules for Training and Inference of segment-anything model for myotube segmentation and stardist model for nuclei segmentation.

   Visualisation of Inferece process in the designed tool:

   If you are interested in the Inference Tool, please contact us.
   

   

4. Project Report Contains the Final Report of the project.

5. App Backend Contains the Backend of the Web Application.

6. App Frontend Contains the Frontend of the Web Application.

MyoSAM Model

• MyoSAM checkpoint

Maintainer

{
    name = "Giorgi Nozadze",
    email = "[email protected]"
}

Acknowledgements

We would like to thank:

• Our Supervisors and Project Partner for their support and guidance throughout the project.

• Musculoskeletal Center at the LMU University Hospital for providing the necessary infrastructure and resources for the project.

• Meta AI for their open-source research and materials Segment Anything. We used their work for instance segmentation of myotube images.

@article{kirillov2023segany,
  title={Segment Anything},
  author={Kirillov, Alexander and Mintun, Eric and Ravi, Nikhila and Mao, Hanzi and Rolland, Chloe and Gustafson, Laura and Xiao, Tete and Whitehead, Spencer and Berg, Alexander C. and Lo, Wan-Yen and Doll{\'a}r, Piotr and Girshick, Ross},
  journal={arXiv:2304.02643},
  year={2023}
}

• Developers of the StarDist project for making their great work avaliablle open-source. We used their model for instance segmentation of nuclei images.

@inproceedings{schmidt2018,
  author    = {Uwe Schmidt and Martin Weigert and Coleman Broaddus and Gene Myers},
  title     = {Cell Detection with Star-Convex Polygons},
  booktitle = {Medical Image Computing and Computer Assisted Intervention - {MICCAI}
  2018 - 21st International Conference, Granada, Spain, September 16-20, 2018, Proceedings, Part {II}},
  pages     = {265--273},
  year      = {2018},
  doi       = {10.1007/978-3-030-00934-2_30}
}

@inproceedings{weigert2020,
  author    = {Martin Weigert and Uwe Schmidt and Robert Haase and Ko Sugawara and Gene Myers},
  title     = {Star-convex Polyhedra for 3D Object Detection and Segmentation in Microscopy},
  booktitle = {The IEEE Winter Conference on Applications of Computer Vision (WACV)},
  month     = {March},
  year      = {2020},
  doi       = {10.1109/WACV45572.2020.9093435}
}