My parents sent me a breathing exercise video by complaining that they could not follow the physiotherapist. They needed a simpler way to perform such exercises in order to strengthen their lungs. After that day, I have seen a few more voluntary doctors and nurses trying to help people by posting breathing exercise videos. By completing a researching on the topic, I was convinced that these exercises are not only to be more healthy but vital to survive in today's circumstances. The gravity of these exercises comes from the characteristic of the virus itself. The virus infects the lung cells and an excess of inflammation liquid comes out. This inflammation liquid prevents O2 transfer to blood cells and patient starts to have breathing problems. The current treatment is to perform breathing exercises to get rid of the inflammation liquid for early stages.
What it does
So, we developed a guide for the user to perform such breathing exercises in a more convenient and healthy way. The idea was to come with a web app to perform breathing exercises to the user. The app checks if the user is doing the breathing exercises right. The web application is developed using python for back-end and angular for front-end.For the app development we used angular9 framework and libraries such as opencv, flask and sklearn for computer vision part.
How we built it
We started with exploring the image recognition tools that we can use. Our first idea was to build and mobile application, so we looked for lightweight solutions. After investing some time we decided to go with web application because of its flexibility, limited time and bigger community for support in case we need it.
Challenges we ran into
The main challenge was to track an unlabeled object in a very detailed way. Since the chest or stomach movements are barely observable during respiration by bare eyes it made our case even more challenging to track such movement with a single camera. Taking consideration of uncontrolled lighting and the image quality, environment did not help our case so does the virus. However, we enjoyed the challenges and reached a satisfying version considering time and academic resources about the virus.
Accomplishments that we're proud of
We are proud to be able to measure the inflation and deflation of lungs and chest cavity separately. In that way, we were able to perform breathing exercises using both the diaphragm and lungs. By using explicit and visual guidance, we tried to make sure that the user performing the therapy in the healthiest way. The control mechanism which was able to give the user live feedback ensured the most out of the exercise.
What we learned
What's next for BrEx
Many patients nowadays requires extensive care by physiotherapists and their relatives because of the lung related diseases like Covid-19. Day by day the natural, non chemical treatments like this gain popularity and the application area like ours are expanding.
Log in or sign up for Devpost to join the conversation.