Inspiration
This project was inspired by the low number of medical specialist we have in some areas in Africa, particularly in my country, Nigeria. And some other parts of the world especially in the times that demands for a lot of medical professional. The lack of experience of the few available medical professionals and medical tools in some of these areas also cause misdiagnosis.
What it does
MAAS is an open source app that allows Medical practitioners with low level of experience and also patients to access AI models created by Research teams as a very affordable service in order to get faster and safer diagnostics.
The MAAS project consists of a collection of well trained AI models that each specialize in single medical fields to the fingertip of the users, anywhere, anytime. It's like having not just one but many specialists in your pocket without any delay in attending to you.
How we built it
The project began with thorough research to identify medical conditions that could be diagnosed using medical imagery such as CT scans and MRI scans. I then identified available datasets suitable for training AI models. The models were trained using Google Colab, and an Android app was developed using primarily Java. The project is well-organized and hosted on the MAAS GitHub repository. Google Colab and TensorFlow were used to train the AI models, while Streamlit was employed to host and serve the Python script that makes the AI models accessible.
Challenges we ran into
I encountered challenges related to learning and adapting to the technology required for the project's proof of concept.
Accomplishments that we're proud of
One of my notable accomplishments is the development of a functional app in under 18 hours.
What we learned
Through this project, I learned valuable lessons in technology integration and development, as well as the importance of addressing the shortage of medical specialists in underserved areas.
I also learned about the really underrated need of having a team to build something as impactful as this with me. I would have been able to achieve more with more people working on MAAS with me.
What's next for MAAS 1.0
In the future, I and a future team plan to expand and enhance MAAS by adding more AI models to cover a broader range of medical conditions. We also aim to improve the user interface and experience, integrate additional medical imaging types, and potentially collaborate with healthcare institutions for real-world testing and implementation.
APK demo link
Log in or sign up for Devpost to join the conversation.