Inspiration
Limited healthcare access to underprivileged due to lack of financial support and medical infrastructure can be addressed through an AI-enabled robotic telemedicine kiosk. The kiosk provides user-friendly conversational interface to identify medical issues, enables remote consultations with doctors through telemedicine apps. This improves availability and accessibility of primary healthcare for underserved population of the country.
What it does
- Develop a AI interface that can understand common health symptoms and complaints in various regional languages when described by users, in order to bridge the language barrier between patients and doctors.
- Enable remote consultations by integrating telemedicine functionality to connect patients to specialists based on preliminary diagnosis from the kiosk.
How we built it
The kiosk provides user-friendly conversational interface to identify medical issues, enables remote consultations with doctors through telemedicine apps.
We utilized Python and open source frameworks to generate a vector database of 5 different textbooks related to different sub branches of medicine. They cover fields like general medicine, ICU care, Critical treatment. The entire flow is deployed on a Flask application and is called using a website built using basic HTML, CSS and JS.
Challenges we ran into
Currently in the stage where we are running into problems for integration of AI and HTML JS stuff.
Accomplishments that we're proud of
We are proud of the final model we achieved after 5 hours of finetuning. When asked with real doctors about the diagnosis by our AI-Doc, the feedback was positive which made us realize the potential of the project.
What we learned
Apart from the technological knowledge gained in the fields of Generative AI and Web development, the skills necessary in daily life like perseverance to code till the finish of hackathon and the power we discovered in the word team.
What's next for AI assisted Telemedicine KIOSK
We plan to train the model further more using many more textbooks and make the model more reliable and try to deploy it over a cloud server for faster response generation. Apart from that, we also plan to include priority queues for each doctor so that a patient in need can find the available doctor as soon as possible.
Log in or sign up for Devpost to join the conversation.