Inspiration

As the pandemic is growing, patients are worried if they can get a bedspace, PPE kits, Food, etc., One concern about the impact of COVID-19 on hospitals is ICU capacity. It is no secret that at the center of COVID-19 Pandemic there aren’t simply enough ICU beds to support the number of cases. At present, data on real time hospital bed availability is not accessible to public. It is difficult to find a hospital bedspace and what gets even more challenging is the service requests to tend to. Every second counts!

A treatment without hassle and a timely self assessment, can provide confidence to the user and a sense of trust for the community which will make the rest of 2020 easy for us. The main aim of our app is to act fast during these times.

What it does

A self assessment is a set of questionnaire that helps to identity if user has COVID. These include uploading a copy of x-ray scan of lungs to be sure of any Pneumonic symptoms. Azure customvision.ai service used to train a machine learning model with Kaggle dataset of x-ray images. After the assessment, they can visit a nearby testing center to confirm. Once the user is COVID positive, they can act fast by registering a bed-space for themselves at the nearby hospital. API is created that finds nearby hospitals with available bed-space and charges based on user location. After bed-space allocation, they can request for additional resources like food, PPE kits, water etc., from the app itself.

Hospital management consolidates the requests and tend to them in the order they arrive. This keeps hospital management organized and helps the staff and patients to have their needs met with social distancing in practice as every request goes through the app only. Additionally, maintaining social distancing is important for both patients and hospital staff. To ensure that, we made use of Google SODAR which creates augmented reality of two meter radius ring around the user. With this feature staff and patients can move safely inside the hospital.

How we built it

Android - main app Express Js, Node Js - Microservices deployed on Azure as an app service Kaggle datasets - Gathered X-ray images and Hospital-bed information from kaggle Microsoft SQL server management studio - To deploy the datasets onto azure backend Google SODAR - AR functionality for following social distancing Microsoft Azure - Backend cloud for the application

Challenges we ran into

Integrating android with machine learning model developed in customvision.ai achieving an accuracy of ~98% for x-ray classification model

Accomplishments that we are proud of

  1. A real time classification of X-ray images into COVID positive and negative for better judgement.
  2. Integration with Google Sodar to enable safe navigations within hospital
  3. Making the app intuitive

What we learned

Azure Cognitive services - as a part of customvision.ai which enabled us to understand the working of image classification and deploying an end point to it to be able to use it as a microservice.

What's next for CoviCure

  1. map integration for intutive hospital selection
  2. Developing social distancing AR with haptic feedback
  3. Perform an analysis of the healthcare system capacity as compared to disease spread forecasts and efficiently plan resources.
  4. Use of "Notification Hubs" service from Microsoft Azure to manage the status of requests within the app
  5. Using load balancers for application scaling

Built With

Share this project:

Updates