Skip to content

anantp1910/WatchCare-AI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

WatchCare-AI

Real-time fall detection and alerting system designed for hospital waiting rooms to prevent patient incidents from going unnoticed in understaffed emergency environments.

Built at Carnegie Mellon University's NexHacks (24-hour hackathon).


🚨 Problem

Hospital ER waiting rooms are often understaffed.

Patients may:

  • Slump in chairs
  • Faint while seated
  • Slide off chairs
  • Collapse without immediate supervision

Traditional camera monitoring requires constant human attention. In busy environments, subtle fall-related incidents can go unnoticed.

WatchCare-AI provides automated real-time monitoring and alerting to reduce response time and improve patient safety.


⚡ Solution Overview

WatchCare-AI is a computer vision pipeline that:

  1. Monitors live video feed
  2. Performs real-time pose estimation
  3. Detects abnormal body angles and motion persistence
  4. Triggers instant nurse alerts
  5. Enables claim-and-resolve tracking for accountability

The system is designed to minimize false positives in safety-critical environments.


🧠 System Architecture

🔍 Computer Vision Layer

  • Python
  • OpenCV
  • Pose estimation model
  • Body angle threshold detection
  • Motion persistence filtering logic

The detection logic evaluates:

  • Torso angle deviation
  • Sudden vertical displacement
  • Sustained abnormal posture duration

This prevents alerts from being triggered by normal seated movements.


🔔 Alerting Layer

  • Firebase Firestore
  • Real-time event synchronization
  • Audit logging
  • Claim-and-resolve workflow for nurse accountability

Alerts include:

  • Timestamp
  • Camera ID
  • Incident classification
  • Status tracking (Open / Claimed / Resolved)

📱 Mobile Integration

Collaborated with Android developers to deliver:

  • Instant push notifications
  • Real-time alert dashboard
  • Ownership tracking
  • Resolution confirmation

🏗 Design Principles

  • Real-time responsiveness
  • False-positive minimization
  • Clear ownership assignment
  • Scalable system design
  • Audit-ready logging

🛠 Tech Stack

  • Python
  • OpenCV
  • Pose Estimation
  • Firebase Firestore
  • Android (integration layer)

🧪 Detection Logic (Simplified)

The system evaluates:

  • Body angle > predefined threshold
  • Vertical drop speed exceeding tolerance
  • Motion persistence over defined time window

Only if multiple conditions are satisfied is an alert triggered.

This layered validation reduces alert noise in busy hospital environments.


⏱ Built In

Developed in a 24-hour hackathon environment at Carnegie Mellon University's NexHacks.

Focused on:

  • Real-world healthcare safety gaps
  • Scalable architecture
  • Deployable MVP logic

📈 Impact Potential

  • Reduced ER monitoring burden
  • Faster incident response time
  • Improved patient safety
  • Real-time accountability tracking

🔮 Future Improvements

  • Edge-device deployment for latency reduction
  • HIPAA-compliant cloud architecture
  • Multi-camera synchronization
  • Fall severity scoring
  • Machine learning refinement using real incident datasets
  • Integration with hospital paging systems
  • Privacy-preserving anonymized pose detection

👩‍💻 Contributors

Built by:

  • Varnika Yadav
  • Devaj Solanki
  • Anant Patel
  • Anusha Agarwal

📌 Why This Matters

WatchCare-AI demonstrates applied computer vision, real-time systems engineering, and safety-critical system design under time constraints.

It showcases practical AI deployment beyond theoretical modeling.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors