Skip to content

Latest commit

 

History

History
88 lines (65 loc) · 3.39 KB

README.md

File metadata and controls

88 lines (65 loc) · 3.39 KB
Raw

WMHSegmentation-Thesis-Project

Table of Contents

  1. Introduction
    • Overview and explanation of the project.
  2. Objectives
    • Development, automation, and validation of the lung image registration process.
  3. Folders
    • Descriptions of various folders like Noteebooks, Data, Parameters, etc.
  4. Installation and Usage
    • Software requirements, installation guide, and usage instructions.
  5. How to use it
    • Links to instructons on how to use the model.
  6. Authors
    • Contributions and profiles of the project team.
  7. License
    • Licensing information of the project.

Introduction

This repository houses the project of the thesisAutomated Segmentation of White Matter Hyperintensities using Deep Learning , a crucial task in the early diagnosis and intervention of neurodegenerative diseases like Alzheimer’s. Our method uniquely integrates 3 architecture with multi-planar data representation and innovative training techniques to enhance segmentation accuracy and robustness. Implemented models in this code:

alt text

Objectives

  1. Comprehensive analysis of medical image analysis for White Matter Hyperintensities (WMH).
  2. Evaluate the performance of Dense UNet, FastSurferCNN, and nn-UNet.
  3. Investigate the effects of multimodal information and varying input types and kernel sizes on these models.
  4. Explore the impact of transfer learning.
  5. Explore the models Generalizability.

Folders

  • Config: configurations files for the networks and default configuration.
  • Data : data scripts for ingesting, altering and preprocessing.
  • Metrics : Evaluation Metrics use for the models.
  • Models : Pices of code use in the creation of the models.
  • Utils : General codes use in the model.
  • Examples : how to run in cmd line.

Installation and Usage

Prerequisites

Ensure you have Python 3.8 or higher installed, along with PyTorch 1.7 and other necessary libraries detailed in the requirements.txt.

Creating a Virtual Environment

To avoid conflicts with other Python projects, it's recommended to create a virtual environment:

  1. Install virtualenv if you haven't already: pip install virtualenv
  2. Create a new virtual environment: virtualenv venv (or python -m venv venv if using Python's built-in venv)
  3. Activate the virtual environment:
    • On Windows: venv\Scripts\activate
    • On macOS and Linux: source venv/bin/activate
  4. Your command prompt should now show the name of the activated environment.

Installation

Clone the repository using:

git clone https://github.com/EdAlita/WMHSegmentation-DenseUNet.git
cd WMHSegmentation-DenseUNet

Install the required Python packages:

pip install -r requirements.txt

Wiki

How to used the scripts

Authors

License

This project is licensed under the Creative Common Lincense - see the LICENSE.md file for details.