Inspiration

Like most people in the world right now, we are genuinely concerned about COVID-19. One of the biggest problems with this new virus is the fact that it's impossible to know if you are infected without a test, and that “not knowing” is what makes this situation so scary for most of the people. COVID-19 tests are difficult to find because they are simply not enough and cannot be manufactured quickly enough, which is causing panic. Due to this fact, we will have to rely on other diagnostic techniques.

In order to solve this problem we thought about using X-Ray images as doctors frequently use X-rays and CT scans to diagnose pneumonia, lung inflammation, abscesses, etc. Since this virus attacks our respiratory tract, we can use X-rays to analyze the health of a patient’s lungs. And given that most of the hospitals have X-Ray imaging machines, it could be possible to use X-rays to diagnose COVID-19 without the dedicated test kits. A disadvantage is that the analysis of X-Ray images requires a radiology expert and takes significant time.

Here comes our possible solution namely the development of an automated analysis system based on artificial intelligence that can save medical professionals valuable time.

What it does

The system consists of an application in which the doctor uploads the patient's X-ray and the artificial intelligence algorithm from behind returns a percentage that refers to the patient's condition (infected or healthy).

The website also shows realtime global statistics about the pandemic and provides useful information for prevention and protection against the virus.

How we built it

Front-end: HTML, CSS, Javascript, Boostrap Back-end: PHP, Python Machine learning: Tensorflow and Keras frameworks

Challenges we ran into

  • deployment of the AI model on the web

Accomplishments that we are proud of

  • obtaining a 99% accuracy of the AI model

What's next for COVID-Z

  • improvement of the AI model
  • publishing a full version online for free use
  • promoting our solution
Share this project:

Updates