AI-powered system that not only detects falls from CCTV cameras,
but also predicts fall risk in advance to enable early intervention
and faster emergency response.

• Real-time fall detection using AI
• Pose detection with YOLOv11n-pose
• Fall classification using XGBoost
• Risk prediction using mobility trend analysis
• Caregiver notification system
• EMS coordination dashboard
• Real-time location tracking with map
• Designed for CCTV and Edge AI devices

The system analyzes video from CCTV cameras to detect and predict fall incidents.

1. Detect human pose and movement
2. Analyze mobility trends (walking speed, body sway)
3. Classify fall events
4. Send alert to caregiver
5. Caregiver reviews video
6. Forward to EMS if needed
7. EMS dispatches response

Fall_Detect/
├── ai/        # AI models, training, feature extraction
├── edge/      # Real-time detection and alert system
├── backend/   # APIs and Firebase integration
├── web/       # Caregiver & EMS dashboards (Next.js)
├── models/    # Trained models
└── requirements.txt

• Python
• YOLOv11n-pose
• XGBoost
• OpenCV
• Computer Vision
• Machine Learning

• Edge AI
• REST API

• Next.js
• React

• Firebase

• Google Maps API

• Accuracy: ~91%
• Precision: ~88%
• Recall: ~95%

The model is optimized to prioritize recall, ensuring reliable detection
and minimizing missed fall incidents in real-world scenarios.

• Video processing is handled on edge devices
• Only essential data is transmitted
• Designed for real-time emergency response
• Ensures both privacy and reliability

This system improves safety for elderly people living alone by reducing
response time in emergencies. By predicting risks in advance, it enables
early intervention and can potentially save lives.

• Goes beyond detection → predicts fall risk
• Combines AI, alerts, and EMS in one system
• Uses Edge AI for real-time processing and privacy

• Elderly people living alone
• Caregivers and family members
• Emergency response teams (EMS)
• Healthcare providers

• Early fall prediction (before fall happens)
• Improve model accuracy with more data
• Integrate wearable sensors
• Expand to other health emergencies

git clone https://github.com/Fewdev1122/intelligent-Fall-Detect.git
cd intelligent-Fall-Detect
pip install -r requirements.txt
python -m edge.camera

From fall detection → risk prediction → emergency response, this system transforms reactive monitoring into proactive safety.