Explainable Detection of COVID-19 from Chest X-Ray Images

This project is part of the Deep Learning in Data Science course (DD2424) at KTH. The goal is to train a classifier for COVID-19 detection from chest X-ray (CXR) images and boost it with explainability. More information can be found in the report. Also check out our presentation.

This package provides:

• An application with a graphical user interface (GUI). This application can be used to make predictions on your images using trained models.
• A suite of tools with a command-line interface (CLI). These tools can be used to train and test new models.
• Several modules with a Keras-like API. These modules can be used in Python code.

1. Setup

The recommended installation is the following:

wget https://raw.githubusercontent.com/franco-ruggeri/dd2424-covid19-detection/master/scripts/install.sh -O install.sh
bash -i install.sh

Following the prompt, you can get a ready-to-use installation that uses the best models we trained.

The package is distributed on PyPi, so can be installed also with:

pip install covid19-detection

However, in this case you have to provide the trained models to the application. You can decide either to download the best models we trained or to train your own models with the command-line tools.

2. Application

If you have done the recommended installation, you can launch the application by searching it among the applications. Otherwise, you can launch it from the terminal:

covid19-detector

3. Command-line suite

The command-line suite is available under the covid19-detection command. It provides several subcommands. The list can be retrieved with:

covid19-detection -h

More information about each subcommand can be obtained with:

covid19-detection <subcommand> -h

4. Package

You can import the package in your Python code with:

import covid19

The covid19 package is composed of the following sub-packages:

• covid19.datasets: contains utilities for generating COVIDx, HAM10000 and for building an input pipeline with tf.data.
• covid19.models: contains ResNet50 and COVID-Net, two deep convolutional neural networks.
• covid19.explainers: contains Grad-CAM and IG, two explainable AI methods, with some utilities for plotting the explanations.
• covid19.layers: contains layers used by models in covid19.models.
• covid19.metrics: contains utilities for computing and plotting metrics.
• covid19.gui: contains graphical user interface implemented with Qt.
• covid19.cli: contains command-line interface.

Each subpackage provides interesting modules. For example, you can create a COVID-Net as follows:

from covid19.models import COVIDNet

model = COVIDNet(n_classes=3)

For more information about each class, see the comments.