Inspiration
Tuberculosis is the world's deadliest infectious disease and is underprioritized in developing countries such as India or the Phillipines. One of the best ways to treat it is early diagnosis to prevent serious illnesses and the continued spread of the disease between people. By creating a simple speech-to-text web app that anyone can complete at home, we can help catch more cases of TB early and make it easier for doctors to treat patients.
What it does
TB Screener uses Vertex AI Gemini Pro enhanced with RAG (Retrieval-Augmented Generation) on culture-specific tuberculosis medical literature to prompt users to complete a self-screening. It uses speech-to-text input, allowing people who speak different languages or are unfamiliar with typing to fill out the pre-screening. Generative AI then summarizes the patient's comments and gives some technical suggestions for the doctor in a formatted PDF report.
How we built it
Gemini Pro Vision w/ RAG -> Python and Flask -> React Frontend Not sleeping
What we learned
We learned a lot about the abilities and limitations of generative AI!
Log in or sign up for Devpost to join the conversation.