This project focuses on detecting human posture using OpenCV and MediaPipe. Maintaining good posture is essential for physical and mental well-being, as poor posture can lead to chronic discomfort, mobility issues, and musculoskeletal disorders.
With an increasing number of software developers and professionals working long hours at desks, posture-related problems are becoming more prevalent. This project aims to analyze and determine whether a person's posture is good or bad based on body angles calculated from tracked distance vectors.
- Posture Classification: Determines whether the detected posture is good or bad.
- Real-time Pose Estimation: Uses OpenCV and MediaPipe to track human body joints.
- Angle Calculation: Measures angles between key body joints to assess posture.
- Lightweight & Efficient: Utilizes pre-trained models for fast and accurate results.
- Python
- OpenCV - Open-source computer vision and machine learning library.
- MediaPipe - Google's framework for real-time pose estimation.
- Pose Detection: MediaPipe collects 33 key body points such as shoulders, elbows, etc.
- Angle Calculation: Computes angles between body joints to analyze posture.
- Posture Classification: Determines whether the posture is correct or incorrect.
git clone https://github.com/txrunteja/human-posture-detection.git
cd human-posture-detection