Covid Care

Test image for Pneumonia X-ray
Test image forNormal X-ray
Test image for COVID-19 X-ray

Inspiration

The pandemic has changed our lifestyle and until today it is affecting us, and it is necessary to know every information to keep us and our loved ones safe. With the virus mutating and increased cases of asymptomatic cases now COVID can be being detected using the chest x-rays as it affects the respiratory system first. According to LSU radiologists, these chest x-rays can be used for rapid diagnosis where there are limited testing and delayed results. And as we all know prevention is better than cure with a lot of advice from everyone there is a lot of confusion over the right and spam news. There is a need for expert advice, current news, and rapid diagnosis.

Our Solution

Virtual communication is the new future. In order to help the nurses, patients, and doctors in an effective way, we have created an application that predicts whether a person has COVID or not using Xray inputs from the user and giving advice on the results, The application also gives all the latest information on COVID-19, the preventions following the CDC guidelines, the statistics of COVID-19 cases, and an easy platform to contact doctors to seek advice virtually.

How we built it

We used cloud technologies for our applications The model used for COVID prediction was developed using Google’s vision API which achieved a 91% accuracy. Then we designed the front end of the application using Vue-js and HTML. The application is hosted using Google's firebase using its cloud function and user authentication services.

Challenges we ran into

The main challenge was to collect the different data set and train to get the model as we had to perform multiclass classification with high accuracy. The model was deployed in the cloud and was used by the application by REST Api but since it was developed using Vue we faced issues to convert the image and pass the input. And lastly, we faced a little difficulty in time management as we developed the model from scratch and it takes several hours.

What's next for Covid Care

This application can be used for a complete record maintenance and session management for doctors and adding 24/7 chat bot service for emergency consultation.

Built With

google-firebase
google-vision
html
python
vue

Submitted to

ShellHacks 2020
- Winner Best Use of Google Cloud (COVID-19 Hackathon Fund) by Google
- Winner Wolfram Award by Wolfram Language
- Winner Best Use of Google Cloud by Google - 2nd Place

Created by

I worked on building the model for detection of Covid-19 . I collected the data from various sources and feed those images to Google cloud Auto ML vision API to create the model.

Harsh Kathiriya
I helped the team with UI in vue js and also integrating frontend with backend using Firebase.

parthavi96 Kaka
I worked in building the firebase authentication and cloud functions, also I helped in integrating the cloud vision API

Harshit Ambalkar
Worked on collecting different datasets and combining them and developing the model using Google Vision API.

Ushasree Boora