SafeCare AI – Smart Patient Monitoring System

Privacy-aware, AI-powered system for real-time detection of patient falls and wandering using video analytics.

Overview

Hospitals face staff shortages, and continuous patient monitoring is difficult. SafeCare AI uses real-time pose estimation to automatically detect falls, wandering, and risky movements, sending instant alerts to nurses — while protecting patient privacy through on-device processing and anonymized video.

Features

• Real-time fall detection using pose estimation (MediaPipe)
• Bed-exit / wandering detection with virtual "bed zone"
• Anonymized video (blur + skeleton overlay)
• React dashboard with:
  • Live video feed
  • Alerts panel (fall / wandering)
  • Popup notifications with room info
• Automatic alert popup + sound on new incident

Tech Stack

• Frontend: React, JavaScript, CSS
• Backend: Python, FastAPI
• AI/ML: OpenCV, MediaPipe Pose
• Others: In-memory event store, JavaScript Audio API

Project Structure

• backend/ - FastAPI server, video processing, AI logic
• frontend/ - React dashboard (live view + alerts)

Running Locally

Backend

cd backend
python -m venv venv
venv\Scripts\activate  # Windows
pip install -r requirements.txt
uvicorn app:app --reload --port 8000

Frontend

cd frontend
npm install
npm run dev