Overview: An AI-powered vision system designed to monitor workstation ergonomics in real-time, notifying the user when poor posture is detected to prevent musculoskeletal issues.

1. The Integration of Two Worlds:

• As a Physical Activity Technician, I identified the need for an automated tool to correct sedentary posture.
• I developed a custom solution using Computer Vision to track key body landmarks (shoulders, neck, spine) during work sessions.

2. Technical Stack:

**Language:** Python.

• Libraries: Mediapipe or OpenPose for skeleton tracking, OpenCV for real-time video processing.
• Logic: Implementation of angular thresholds to trigger alerts when the head tilts forward or shoulders slouch beyond ergonomic limits.

3. Key Features:

• Real-time Inference: Low-latency processing optimized for background execution.
• Smart Notifications: Visual or auditory alerts that trigger only after a sustained period of bad posture to avoid "alert fatigue."
• Privacy-First: Designed to process frames locally without storing sensitive video data.

PostureGuard AI - Computer Vision for Ergonomics