Icarus MC for ART

Inspiration

Hello Everyone, we are Icarus from NTU and we are a bunch of aspiring Software Engineers. In Singapore, the number of COVID-19 cases are rising sharply, to record highs almost daily. This brings about a large strain on the healthcare system and in particular, healthcare workers. We noticed that there is a need for healthcare workers to have more capacity for important cases such as life-threatening situations instead of cases that are low risk or with no symptoms. This project is meant to fill this niche and automate processes in order to alleviate pressure on their already immense workload.

What It Does

We thought of a simple idea to help Singapore's doctors issue Medical Certificates (MC) through the use of a Telegram Bot alongside a Cloud Service. Check our Github Repo for more details.

How We Built it

We built a Convolutional Neural Network to help with Image Classification and through the use of keras/tensorflow and python scripting, we predict on an image fed to us to determine the status of the submission (i.e. whether to give an MC or not.)

Challenges We Ran Into

Lack of data for training, poor data quality, authentication, Telegram and cloud incompatibilities. We attempted to recreate ART kit data by populating the dataset with images that have been modified in various ways to increase the robustness of our model.

Accomplishments that We're Proud Of!

We managed to apply software engineering methodologies such as utilizing an agile workflow with a scrum methodology. We have learnt new technologies and did our best to implement them in this proof of concept within the given time.

What We Learned

Time under pressure = growth. Data processing is the biggest part of any Machine Learning/AI/Software Engineering product.
Solution has very real potential with regards to cloud computing and good scalability in terms of getting a better dataset, using alternative models/parameters and lastly, integration with already existing infrastructure, such as Singpass and TraceTogether.
We hope that our solution would be able to bring in fresh insights and inspire other possible solutions to this very pressing issue in Singapore.