Inspiration

Doctor visits are time consuming and repetitive. During checkups, patients are expected to wait for long periods of times for simple diagnosis while they are forced to pay large amounts for just a checkup. We hope to solve this problem with our IOS native app, Salutem.

One of our member's grandpa arrived in America without health care and needed help instantly. Due to the lack of communication and availability of doctors in the area, he was put in a hard situation. Hence, we were encouraged to design a product that not only lowers wait times at the clinic but also makes everything more efficient providing on the spot skin and eye checkups and sending those results instantly to the doctors close to you.

What it does

To mitigate this struggle, we created Salutem to allow for easy and accurate machine learning-based diagnoses, paired with expert consultation from our vetted doctors. Together a user can perform a simple user-based check-up and test for various diseases, and then, they can quickly relay this information to the doctor if there is an impending danger. Additionally, we allow for doctors to prescribe medications and offer any further support through our chat-based environment in the app as well as through email.

How I built it

Our model leverages a variety of cutting edge technologies to develop the best user experience and the most effective inappropriate diagnosis available in the current market. Our Machine Learning algorithms are based on the Keras API and a Tensorflow Backend which was hosted on a cloud computing service for quick data analysis.

In addition, our app uses MapKit to determine our patient's locations on an active and accessible grid. The app development was created in Xcode using Swift. In order to communicate between patient and doctors, we employed a Firebase backend to easily store information and custom IDs about our individual patients This technology was further enhanced to allow for quick sign in and sign off. Finally, we used JSON Mapping tools tp analyze different nodes in the parent hierarchy. All in all, these technologies allowed for the development of a one of a kind, cutting edge app aimed to solve today's greatest health challenges!

Challenges I ran into

Some challenges I ran into were some of the database queries. Our team and I had difficulty parsing through the hierarchy of the database, causing some challenges for us.

Accomplishments that I'm proud of

Some accomplishments that I'm proud of is implementing the neural network in the mobile application itself. Exporting .model files from a neural network using a local CPU is quite a difficult task, however, by implementing a better model architecture, this task was simplified, and we were proud of this!

What I learned

During our Hackathon, we learnt several different technologies as well as furthered our skillsets in several relevant APIs such as Tensorflow and Keras. Additionally, we learnt how to integrate a Firebase backend into an IOS native app. Further, we learned how to use the Market to display professional-looking maps, target locations, and relevant data. The essence of our project is based on the new technologies offered to us, which would have otherwise been foreign to us.

What's next for Salutem

This technology can be leveraged to work with several other skin/surface diseases and diseases which can be understood through other kinds of image and NLP data. In the future, we plan to further develop our app by adding more custom features, showcasing detailed summary reports of diseases, and using more scientific and accurate data. Additionally, we plan to use sentiment analysis and chatbot algorithms to provide machine learning-based solutions to various diseases and ultimately remove the need for a doctor diagnosis!

Link for slides: https://docs.google.com/presentation/d/1uaPM_gdHRUpDQnijz8bGsEFDpMlSYKLO8duzkmipHF0/edit?usp=sharing

Share this project:

Updates