Baseline inference Algorithm for the STOIC2021 challenge.

Overview

STOIC2021 Baseline Algorithm

This codebase contains an example submission for the STOIC2021 COVID-19 AI Challenge. As a baseline algorithm, it implements a simple evaluation pipeline for an I3D model that was trained on the STOIC2021 training data. You can use this repo as a template for your submission to the Qualification phase of the STOIC2021 challenge.

If something does not work for you, please do not hesitate to contact us or add a post in the forum. If the problem is related to the code of this repository, please create a new issue on GitHub.

Table of Contents

Before implementing your own algorithm with this template, we recommend to first upload a grand-challenge.org Algorithm based on the unaltered template by following these steps:

Afterwards, you can easily implement your own algorithm, by altering this template and updating the Algorithm you created on grand-challenge.org.

Prerequisites

We recommend using this repository on Linux. If you are using Windows, we recommend installing Windows Subsystem for Linux (WSL). Please watch the official tutorial by Microsoft for installing WSL 2 with GPU support.

  • Have Docker installed.
  • Have an account on grand-challenge.org and make sure that you are a verified user there.

Building, testing, and exporting your container

Building

To test if your system is set up correctly, you can run ./build.sh (Linux) or ./build.bat (Windows), that simply implement this command:

docker build -t stoicalgorithm .

Please note that the next step (testing the container) also runs a build, so this step is not necessary if you are certain that everything is set up correctly.

Testing

To test if the docker container works as expected, test.sh/test.bat will build the container and run it on images provided in the ./test/ folder. It will then check the results (.json files produced by your algorithm) against the .json files in ./test/.

If the tests run successfully, you will see Tests successfully passed....

Note: If you do not have a GPU available on your system, remove the --gpus all flag in test.sh/test.bat to run the test. Note: When you implemented your own algorithm using this template, please update the the .json files in ./test/ according to the output of your algorithm before running test.sh/test.bat.

Exporting

Run export.sh/export.bat to save the docker image to ./STOICAlgorithm.tar.gz. This script runs build.sh/build.bat as well as the following command: docker save stoicalgorithm | gzip -c > STOICAlgorithm.tar.gz

Creating an Algorithm on grand-challenge.org

After building, testing, and exporting your container, you are ready to create an Algorithm on grand-challenge.org. Note that there is no need to alter the algorithm implemented in this baseline repository to start this step. Once you have created an Algorithm on grand-challenge.org, you can later upload new docker containers to that same Algorithm as many times as you wish.

You can create an Algorithm by following this link. Some important fields are:

  • Please choose a Title and Description for your algorithm;
  • Enter CT at Modalities and Lung (Thorax) at Structures;
  • Select a logo to represent your algorithm (preferably square image);
  • For the interfaces of the algorithm, please select CT Image as Inputs, and as Outputs select both Probability COVID-19 and Probability Severe COVID-19;
  • Choose Viewer CIRRUS Core (Public) as a Workstation;
  • At the bottom of the page, indicate that you would like your Docker image to use GPU and how much memory it needs. After filling in the form, click the "Save" button at the bottom of the page to create your Algorithm.

Uploading your container to your Algorithm

Uploading manually

You have now built, tested, and exported your container and created an Algorithm on grand-challenge.org. To upload your container to your Algorithm, go to "Containers" on the page for your Algorithm on grand-challenge.org. Click on "upload a Container" button, and upload your .tar.gz file. You can later update your container by uploading a new .tar.gz file.

Linking a GitHub repo

Instead of uploading the .tar.gz file directly, you can also link your GitHub repo. Once your repo is linked, grand-challenge.org will automatically build the docker image for you, and add the updated container to your Algorithm.

  • First, click "Link Github Repo". You will then see a dropdown box, where your Github repo is listed only if it has the Grand-Challenge app already installed. Usually this is not the case to begin with, so you should click on "link a new Github Repo". This will guide you through the installation of the Grand-challenge app in your repository.
  • After the installation of the app in your repository is complete you should be automatically returned to the Grand Challenge page, where you will find your repository now in the dropdown list (In the case you are not automatically returned to the same page you can find your algorithm and click "Link Github Repo" again). Select your repository from the dropdown list and click "Save".
  • Finally, you need to tag your repository, this will trigger Grand-Challenge to start building the docker container.

Make sure your container is Active

Please note that it can take a while until the container becomes active (The status will change from "Ready: False" to "Active") after uploading it, or after linking your Github repo. Check back later or refresh the URL after some time.

Submitting to the STOIC2021 Qualification phase

With your Algorithm online, you are ready to submit to the STOIC2021 Qualification Leaderboard. On https://stoic2021.grand-challenge.org/, navigate to the "Submit" tab. Navigate to the "Qualification" tab, and select your Algorithm from the drop down list. You can optionally leave a comment with your submission.

Note that, depending on the availability of compute nodes on grand-challenge.org, it may take some time before the evaluation of your Algorithm finishes and its results can be found on the Leaderboard.

Implementing your own algorithm

You can implement your own solution by editing the predict function in ./process.py. Any additional imported packages should be added to ./requirements.txt, and any additional files and folders you add should be explicitly copied in the ./Dockerfile. See ./requirements.txt and ./Dockerfile for examples. To update your algorithm, you can simply test and export your new Docker container, after which you can upload it to your Algorithm. Once your new container is Active, you can resubmit your Algorithm.

Please note that your container will not have access to the internet when executing on grand-challenge.org, so all model weights must be present in your container image. You can test this locally using the --network=none option of docker run.

Good luck with the STOIC2021 COVID-19 AI Challenge!

Tip: Running your algorithm on a test folder:

Once you validated that the algorithm works as expected in the Testing step, you might want to simply run the algorithm on the test folder and check the output .json files for yourself. If you are on a native Linux system you will need to create a results folder that the docker container can write to as follows (WSL users can skip this step).

mkdir ./results
chmod 777 ./results

To write the output of the algorithm to the results folder use the following command:

docker run --rm --memory=11g -v ./test:/input/ -v ./results:/output/ STOICAlgorithm
Owner
Luuk Boulogne
Luuk Boulogne
GrabGpu_py: a scripts for grab gpu when gpu is free

GrabGpu_py a scripts for grab gpu when gpu is free. WaitCondition: gpu_memory

tianyuluan 3 Jun 18, 2022
Keras-tensorflow implementation of Fully Convolutional Networks for Semantic Segmentation(Unfinished)

Keras-FCN Fully convolutional networks and semantic segmentation with Keras. Models Models are found in models.py, and include ResNet and DenseNet bas

645 Dec 29, 2022
QHack—the quantum machine learning hackathon

Official repo for QHack—the quantum machine learning hackathon

Xanadu 72 Dec 21, 2022
Generating Anime Images by Implementing Deep Convolutional Generative Adversarial Networks paper

AnimeGAN - Deep Convolutional Generative Adverserial Network PyTorch implementation of DCGAN introduced in the paper: Unsupervised Representation Lear

Rohit Kukreja 23 Jul 21, 2022
Geometry-Aware Learning of Maps for Camera Localization (CVPR2018)

Geometry-Aware Learning of Maps for Camera Localization This is the PyTorch implementation of our CVPR 2018 paper "Geometry-Aware Learning of Maps for

NVIDIA Research Projects 321 Nov 26, 2022
SeMask: Semantically Masked Transformers for Semantic Segmentation.

SeMask: Semantically Masked Transformers Jitesh Jain, Anukriti Singh, Nikita Orlov, Zilong Huang, Jiachen Li, Steven Walton, Humphrey Shi This repo co

Picsart AI Research (PAIR) 186 Dec 30, 2022
Learning multiple gaits of quadruped robot using hierarchical reinforcement learning

Learning multiple gaits of quadruped robot using hierarchical reinforcement learning We propose a method to learn multiple gaits of quadruped robot us

Yunho Kim 17 Dec 11, 2022
A JAX-based research framework for writing differentiable numerical simulators with arbitrary discretizations

jaxdf - JAX-based Discretization Framework Overview | Example | Installation | Documentation ⚠️ This library is still in development. Breaking changes

UCL Biomedical Ultrasound Group 65 Dec 23, 2022
The code uses SegFormer for Semantic Segmentation on Drone Dataset.

SegFormer_Segmentation The code uses SegFormer for Semantic Segmentation on Drone Dataset. The details for the SegFormer can be obtained from the foll

Dr. Sander Ali Khowaja 1 May 08, 2022
[TIP 2021] SADRNet: Self-Aligned Dual Face Regression Networks for Robust 3D Dense Face Alignment and Reconstruction

SADRNet Paper link: SADRNet: Self-Aligned Dual Face Regression Networks for Robust 3D Dense Face Alignment and Reconstruction Requirements python

Multimedia Computing Group, Nanjing University 99 Dec 30, 2022
Updated for TTS(CE) = Also Known as TTN V3. The code requires the first server to be 'ttn' protocol.

Updated Updated for TTS(CE) = Also Known as TTN V3. The code requires the first server to be 'ttn' protocol. Introduction This balenaCloud (previously

Remko 1 Oct 17, 2021
Multi-query Video Retreival

Multi-query Video Retreival

Princeton Visual AI Lab 17 Nov 22, 2022
Optimising chemical reactions using machine learning

Summit Summit is a set of tools for optimising chemical processes. We’ve started by targeting reactions. What is Summit? Currently, reaction optimisat

Sustainable Reaction Engineering Group 75 Dec 14, 2022
Nest - A flexible tool for building and sharing deep learning modules

Nest - A flexible tool for building and sharing deep learning modules Nest is a flexible deep learning module manager, which aims at encouraging code

ZhouYanzhao 41 Oct 10, 2022
The official homepage of the COCO-Stuff dataset.

The COCO-Stuff dataset Holger Caesar, Jasper Uijlings, Vittorio Ferrari Welcome to official homepage of the COCO-Stuff [1] dataset. COCO-Stuff augment

Holger Caesar 715 Dec 31, 2022
DeepMetaHandles: Learning Deformation Meta-Handles of 3D Meshes with Biharmonic Coordinates

DeepMetaHandles (CVPR2021 Oral) [paper] [animations] DeepMetaHandles is a shape deformation technique. It learns a set of meta-handles for each given

Liu Minghua 73 Dec 15, 2022
CATE: Computation-aware Neural Architecture Encoding with Transformers

CATE: Computation-aware Neural Architecture Encoding with Transformers Code for paper: CATE: Computation-aware Neural Architecture Encoding with Trans

16 Dec 27, 2022
Face Mask Detection System built with OpenCV, TensorFlow using Computer Vision concepts

Face mask detection Face Mask Detection System built with OpenCV, TensorFlow using Computer Vision concepts in order to detect face masks in static im

Vaibhav Shukla 1 Oct 27, 2021
Hyperbolic Procrustes Analysis Using Riemannian Geometry

Hyperbolic Procrustes Analysis Using Riemannian Geometry The code in this repository creates the figures presented in this article: Please notice that

Ronen Talmon's Lab 2 Jan 08, 2023
Provide baselines and evaluation metrics of the task: traffic flow prediction

Note: This repo is adpoted from https://github.com/UNIMIBInside/Smart-Mobility-Prediction. Due to technical reasons, I did not fork their code. Introd

Zhangzhi Peng 11 Nov 02, 2022