Skip to content

Master Thesis: Quantifying Trust of Segmented Pathologies in Brain Imaging with little supervision.

License

Notifications You must be signed in to change notification settings

matthaeusheer/uncertify

Repository files navigation

Build Status

This repository holds the code for our paper The OOD Detection Blindspot of Unsupervised Lesion Detection which has been accepted for the MIDL (Medical Imaging with Deep Learning) 2021 conference. This project has been carried out in the the scope of a Masters Thesis at the Computer Vision Lab at ETHZ.

Table of Contents

About The Project

Pleaser refer to the MIDL 2021 publication The OOD Detection Blindspot of Unsupervised Lesion Detection.

Getting Started

This is an example of how you may give instructions on setting up your project locally. To get a local copy up and running follow these simple example steps.

Prerequisites

  • pyenv - to manage your local python versions
  • pipenv - manage python dependencies within a virtual environment
  • cuda - well, the NVIDIA stuff
  • cuDNN - more NVIDIA stuff, deep learning on steroids

Installation

  1. Clone the repo
git clone https://github.com/matthaeusheer/uncertify.git
  1. Install python dependencies via pipenv
cd uncertify
pipenv install
  1. Active virtual environment
pipenv shell

Usage

Please check out the python scripts in the scripts folder and jupyter notebooks in the notebooks folder.

License

Distributed under the MIT License. See LICENSE for more information.

Contributors

Matthäus Heer, Physics Institute, University of Zurich Janis Postels, Computer Vision Lab, ETH Zurich Xiaoran Chen, Computer Vision Lab, ETH Zurich Shadi Albarqouni, Helmholtz AI, Helmholtz Zentrum M ̈unchen,

About

Master Thesis: Quantifying Trust of Segmented Pathologies in Brain Imaging with little supervision.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published