Covid-19 Detection : Based on Chest Lung X-Ray Scans

Introduction:

COVID‑19 is a highly infectious disease caused by the SARS-CoV-2 virus. There is no vaccine for this disease presently and due to its highly infectious nature, there have been millions of cases worldwide. In early 2020, CDC developed the testing kit for diagnosing COVID-19. The test is called RT- PCR. However the test has a long turnaround time. Therefore researchers and radiologists have turned towards deep learning as one of the solutions for dealing with this new disease. This gives us an opportunity of detecting possible COVID-19 infections on chest Xrays and quarantine high risk patients while test results are awaited.
This project aims to train a deep learning model which can detect COVID-19 from lung X rays. The results from the model can help prioritize patients for further RT-PCR tests. In cases of a false negative RT-PCR test, a chest Xray can indicate if the patient needs to get retested.
The project can classify lung Xrays into COVID-19, Pneumonia and Normal categories.

Steps:

1. Collect Chest X ray Scans into “COVID-19”, “Pneumonia” and “Normal” folders.
2. Generate labels (“COVID-19”, “Pneumonia” and “Normal”) for the scans.
3. Perform Data Augmentation on the X ray images.
4. Build a model using DenseNet121 architecture.
5. Train the model and perform classification.

Flowchart:

Output:

Check Out the Project here:

1. Google Colab
2. GitHub

References:

1. DenseNet121 Architecture Documentation
2. Chest X-Ray Scans Dataset

Contributors:

Mridul Chavan