COVID Detection Tool

Symptoms based & Chest X-Ray Scan based COVID Detection Tool for early diagnosis of Corona Virus Disease and a Pre/Post-Recovery Diet Plan to remain safe and healthy during the examination/treatment.

Comment

Inspiration:

COVID disease became the most devastating disease in last 1 century. Millions of people lost their life and this affected all of us some way or the other. Conventional diagnosis methods of COVID were becoming less efficient as the virus kept mutating. Thus, in order to detect the virus as early as possible, we came up with a Deep Learning and Neural Network based approach to detect COVID disease.

What it does:

Our model works in two ways. Firstly, those users who do not have a Chest X-Ray Scan can have a symptoms based analysis. For that purpose, we developed a website which takes symptoms and their severity as an input and detect the extent up to which a user can suffer from the disease. Secondly, for the users who have a Chest X-Ray Scan, we built a Convolutional Neural Network based Deep Learning Model that takes the jpg/png image of the Chest X-Ray Scan and predicts whether the person suffers from COVID or not.

How we built it:

For Symptoms based COVID detection, we developed a website using HTML, CSS and JavaScript. This website takes symptoms and their severity as an input and detect the extent up to which a user can suffer from the disease. This website was hosted on "netlify" platform. For Chest X-Ray Scan based COVID detection, we developed a CNN based Deep Learning Application using Python programming, Tensorflow, WX and Keras.

Challenges we ran into:

The dataset collection for building Chest X-Ray Scan based COVID detection was a major challenge. We used 8000 images from Kaggle and 24000 images from GitHub to train our model. Initial accuracy was close to 70% and to make it more efficient was another major challenge that was achieved by increasing the convolution layers as well as taking maximum of pooling layers in our CNN approach.

Accomplishments that we're proud of:

We were able to build our Python Application as well as Website successfully. Around 92% accuracy was achieved in Symptoms based COVID detection and roughly 95% accuracy was achieved in Chest X-Ray Scan based COVID detection.

What we learned:

We learned how to work on large datasets for image processing as well as integrating a python application with the CNN and Deep Learning model that we developed. We also learned about the concepts of hidden layers, max-pooling and average-pooling as well as SoftMax functions that helped us in building our Deep Learning Model.

What's next for COVID Detection Tool:

We plan to integrate the Chest X-Ray Scan based COVID detection with the website of Symptoms based COVID detection so that both the tools can be accessed at a single source. We have kept our code as Open-Source so that all the enthusiastic developers can code for this project and help us in increasing the efficiency.

Built With

Share this project:

Updates