We are excited to share the latest progress on MedVision, our AI-powered real-time medical assistance system built using Gemini Live Agent technology.
What is MedVision?
MedVision is designed to assist in real-time symptom recognition and emergency triage using live video and voice interaction. The system analyzes visual cues and spoken symptoms to help identify possible medical conditions and guide the next steps.
Recent Improvements
Live Video AI Analysis The system now uses Gemini Live to analyze live camera input and detect visual cues such as: • Hand on chest → possible chest pain / cardiac symptoms • Labored breathing → respiratory distress • Head holding → possible headache or neurological issue • Slumped posture → possible shock or fatigue
Improved UI / UX • Camera framing guides to help users position themselves correctly • Symptom quick-reference buttons (Chest, Breathing, Head, Posture) • Real-time visual feedback indicators showing when AI analysis is active
Stability Improvements • Fixed frontend loading issues • Optimized real-time frame capture • Improved Gemini prompt handling for better accuracy
Upcoming Updates • Emergency detection alerts • Medical recommendation engine • Mobile-friendly interface • Integration with healthcare providers
Vision
MedVision aims to become a smart AI triage assistant that can help identify critical symptoms quickly and guide patients toward appropriate medical care.
Built With • Gemini Live API • Python backend • Modern web frontend • Real-time video processing
Log in or sign up for Devpost to join the conversation.