Skip to content

A deep learning approach for analysing muscle architecture from musculoskeletal ultrasound images

License

Notifications You must be signed in to change notification settings

njcronin/DL_Track

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DL_Track

A deep learning approach for analysing muscle architecture from musculoskeletal ultrasound images.


A new version of this repo and the DL_Track algorithm can be found here


To start using this software on your own computer, first use the setup instructions.

To train your own model, you can add your own images and labels to those that are provided, or you can create your own dataset (labelling instructions can be found here). Then use the 'Model_Training.ipynb' notebook to train your model (NOTE: Use a GPU!).

To use my trained models to analyse your own data, simply run the 'Inference_Single_Image.ipynb' or 'Inference_Video.ipynb' Jupyter notebook to analyse individual images or videos respectively. Note that in your images/videos, the fascicles should be oriented from bottom-left to top-right (see image below). If they are not, set the 'flip' variable in the notebook to 1 instead of 0.

If you don't have access to a GPU or couldn't get the main version working, try using the colab version (instructions here). This is more limited, but most of the functionality is the same.

If you find this work useful, please cite the corresponding paper: https://arxiv.org/abs/2009.04790

DL_Track image

About

A deep learning approach for analysing muscle architecture from musculoskeletal ultrasound images

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published