Skip to content

Latest commit

 

History

History
67 lines (46 loc) · 1.71 KB

README.md

File metadata and controls

67 lines (46 loc) · 1.71 KB
Raw

Brain Segmentation

This repository contains a deep learning based project made to train several architectures on MRI brain segmentation performance. Some of the tested architectures are:

Unet

This is a deep learning based project to segmentate brains of fetuses of MRI.

Requirements

  • tqdm
  • opencv
  • nibabel
  • MedPy
  • Keras
  • TensorFlow
  • Scikit-image

Download the tool

Clone the source code, cd into your desired location

(env_name)$ git clone GIT
(env_name)$ cd brain_segmentation

Install requirements from requirements.txt

(env_name)$ pip install -r requirements.txt

Running the project


Before running the project you will have to access the "data" folder and add the images and masks in the "test" and "train" folders. It's recommended to divide the total amount of images in 80% train, 20% test.

Finally you're ready to execute the project:

python train.py --exp name_of_the_training


If you don't chose a name_of_the_training or you pick an existing one the tool will show an error message

You will have to activate the environment every time you want to run the tool.

Setup

You can create a new virtual environment using the venv command:

python -m venv /path/env_name

This will create a virtual environment called env_name in the directory /path. To activate it, run:

source /path/env_name/bin/activate

The environments name should appear at the beginnig of the shell surrounded by parentheses, like this:

(env_name)$

For further information on how virtual envirionments work, check the python documentation.