This repository contains code for 3D-UNet for cardiac MRI image segmentation tasks.
Main file:
3D_Cardiac_UNet.ipynb
Helper functions:
utilities.py
read_cardiac_MRI_segments.py
Pure_SA_3D_8x8_2.architecture.py
preprocess_on_fly.py
metrics.py
The notebook contains both training code and the inference code.
- Open the notebook and remove all the commands related to Google Colab and Goole Drive
- Adjust the paths to directories containing dataset in your computer accordingly
- Create a Python virtual environment
- Install the dependencies in the file: requirements.txt
- Run all the cells in the notebook
- Open the notebook in Google Colab.
- Upoad all the files in this repository in the '/content' folder using Files tab in Colab notebook.
- Make sure to place the data in the Google Drive and and fix the paths in the notebook accoridingly
- Run all the cells of the notebook.