MedVision – Project Update
I would like to share the latest progress on MedVision, an AI-powered real-time medical assistance system built using Gemini Live technology.
Overview
MedVision is designed to assist with 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 appropriate next steps.
Recent Improvements
Live Video AI Analysis The system now integrates Gemini Live to analyze camera input and recognize important visual cues such as: • Hand on chest indicating potential chest pain or cardiac symptoms • Labored breathing posture suggesting respiratory distress • Holding the head indicating possible headache or neurological concern • Slumped posture which may indicate shock, fatigue, or weakness
User Interface Enhancements • Camera framing guides to help users position themselves correctly during analysis • Symptom reference buttons for quick identification of chest, breathing, head, and posture issues • Visual indicators that show when the AI system is actively analyzing video input
System Stability Improvements • Fixed frontend loading issues • Improved real-time frame capture performance • Enhanced prompt handling to improve AI response accuracy
Upcoming Development
Future updates will include: • Emergency detection alerts • A medical recommendation engine • Mobile-friendly interface improvements • Integration with healthcare service providers
Vision
MedVision aims to become an intelligent AI triage assistant that helps identify critical symptoms quickly and supports patients in seeking appropriate medical care.
Technology Stack • Gemini Live API • Python backend • Modern web frontend • Real-time video processing
Log in or sign up for Devpost to join the conversation.