Skip to content

krekiehn/ML4MedWS2023

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

56 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ML4MedWS2023

Project for ML4Med in WS 2023

If you encounter any errors, problems or have suggestions, please feel free to improve this read me.

Get Started

Change data path

look for "data_dir = r'C:\Users\Nicolai\PycharmProjects\ML4MedWS2023\Data\ConvertedPelvisNiftiDataset\Converted Pelvis Nifti Dataset'"

Run Code

if the vanilla training runs, try out your self different parameters and models and so on. For that please copy "Code/MONAI/your_custom_training.py" and rename it and then customize it.

Data

Original Dataset

https://universe.roboflow.com/ks-fsm9o/pelvis-ap-x-ray/dataset/2

Nifti source by Niklas Koser

https://cloud.rz.uni-kiel.de/index.php/s/WRM33znRQjLcMZo

the data ist not included in this repository, because it is too large. but you can automatically download it by running "Code/GetData/GetData.py" in "Code/GetData"

MONAI

Tutorial Source

https://github.com/Project-MONAI/tutorials/blob/main/3d_segmentation/spleen_segmentation_3d.ipynb

Python Environment

by venv in PyCharm

  1. Open Project in PyCharm
  2. Open Settings -> Project -> Python Interpreter
  3. Add new Python Interpreter img.png
  4. Select "Add Local Interpreter"
  5. Select "New environment using Virtualenv"
  6. Select a Python Version (e.g. 3.11) img_1.png
  7. Now you have a clean new Python Environment for a specific Python version
  8. Install all required packages via 'pip install -q "monai-weekly[gdown, nibabel, tqdm, ignite]"' and 'pip install -q matplotlib SimpleITK pandas scipy'

via Dockerfile

docker build -t ml4med:latest .

docker run --gpus device=1 -v /home/sukin699/code/ML4MedWS2023/Data/ConvertedPelvisNiftiDataset:/code/ML4MedWS2023/Data/ConvertedPelvisNiftiDataset -i -t ml4med:latest bash

About

Project for ML4Med in WS 2023

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published