🧍‍♂️ Fall Detection System 🚨

A real-time fall detection system using YOLOv8, OpenCV, and SMTP-based email alerts.
Detects human falls in video or webcam feeds and notifies via email with a captured frame.

---

✨ Features

• 🎯 Accurate Detection using YOLOv8 (yolov8s or yolov8m)
• 🧍 Person Tracking with unique IDs & position history
• ⚠️ Fall Detection Logic based on:
  1. Aspect Ratio (detect lying down)
  2. Sudden Vertical Motion
  3. Screen Position (proximity to floor)
• 📸 Email Alert with captured image via Brevo SMTP
• 📹 Video File & Live Webcam Modes
• ⚙️ Configurable Thresholds and detection sensitivity

---

🛠️ Requirements

Install dependencies with:

pip install ultralytics opencv-python cvzone numpy

▶️ Usage

📼 Run with default video:

python fall_detection.py

🎥 Run with live webcam:

python fall_detection.py live

---

🎮 Controls

Key	Action
`q`	Quit the program
`s`	Switch video/live mode
`r`	Reset fall detection

---

📧 Email Configuration

Uses Brevo SMTP for sending alerts.

In the script, update:

TO_EMAIL = '[email protected]'  # Replace with actual email

Make sure SMTP settings (username, password, or app password) are correctly configured.

---

📝 Notes

• Compatible with any YOLOv8 model (`yolov8s.pt`, `yolov8m.pt`, etc.)
• Video will restart automatically after ending.
• Detection thresholds are configurable for sensitivity tuning.

---

🤝 Contributing

Pull requests, improvements, and feature additions are welcome!