Vibesens
AI-power health monitor
The problem Vibesens solves
What People Can Use It For
VibeSense is an AI-powered webcam health monitoring system designed to help individuals, professionals, and organizations track vital health parameters using only a camera — no external sensors or wearables needed.
Use Cases:
Remote Health Monitoring: Users can check vital signs like heart rate, breathing, and stress level through their webcam in seconds.
Home Wellness Tracking: Perfect for daily health tracking without visiting a clinic.
🏥 Telemedicine Integration: Doctors can remotely analyze real-time vitals of patients during virtual consultations.
Mental Wellness & Stress Detection: Helps users monitor stress and relaxation levels for better work-life balance.
Educational & Research Use: Ideal for universities or research labs exploring AI-based biomedical analysis.
How It Makes Tasks Easier & Safer
Zero Hardware Cost: Traditional health devices cost $100–$1000; VibeSense only needs a webcam.
Universal Accessibility: Works on laptops, desktops, and mobiles – anywhere, anytime.
Privacy-Focused: All processing happens locally on the user’s device; no data leaves the system.
Real-Time Monitoring: Provides instant, accurate feedback (<100ms response time).
Easy to Use: One-click setup, clean dashboard, and visual analytics for non-technical users.
AI-Powered Safety: Intelligent filters reduce motion noise, ensuring stable and reliable readings.
In short: VibeSense democratizes health monitoring — making it affordable, private, and accessible for everyone through the power of AI.
Challenges we ran into
⚙️ Key Challenges & How I Overcame Them
- Unstable Signal Due to Lighting Conditions
Problem: rPPG signal accuracy dropped drastically in low or uneven lighting.
Solution: Implemented adaptive brightness correction and Kalman filtering to stabilize color-based pulse detection.
- Motion Artifacts & Noise
Problem: Small head movements or camera shakes introduced noise in the heart rate signal.
Solution: Used a Butterworth Bandpass Filter (0.75–4 Hz) and real-time motion compensation to clean the data stream.
- Performance Optimization
Problem: Real-time AI processing caused high CPU usage (>60%) initially.
Solution: Optimized OpenCV pipelines, implemented multi-threading, and reduced redundant computations — achieving <25% CPU usage and 30 FPS performance.
- Cross-Platform Compatibility
Problem: Different browser camera APIs and hardware variations caused inconsistent results.
Solution: Standardized video capture using WebRTC and added adaptive frame rate handling.
- UI/UX Design Integration
Problem: Combining real-time video feeds with a 3D glassmorphism UI created lag and visual clutter.
Solution: Used React.js + TailwindCSS optimization and lazy rendering for smooth user experience.
- Data Privacy & Local Processing
Problem: Handling sensitive health data without cloud processing was challenging.
Solution: Designed a fully local AI pipeline ensuring data never leaves the user’s system, aligning with GDPR principles.