Real-Time Posture Checker

A real-time posture detection app built with OpenCV, MediaPipe, and PyQT.
It uses your webcam to detect your body landmarks and determine whether you are sitting or standing with good posture.

---

Features

• Live camera feed with pose landmark tracking
• Calculates a posture ratio using the position of your nose and shoulders
• Displays “Good” or “Bad” posture status in real time
• UI made with PyQT with adjustable settings

---

How It Works

1. MediaPipe Pose detects 33 key body landmarks in real time.
2. The app measures the vertical distance between your nose and shoulders,
   normalized by the distance between your shoulders.
3. This gives a posture ratio that adjusts for distance from the camera.
4. If the ratio falls below a set threshold (default 0.75), it’s labeled “Bad Posture.”

---

Installation

1. Clone the repository

git clone https://github.com/Jshepherd06/Posture-Checker.git
cd posture-checker

2. Install Dependencies

If you are using the provided pyproject.toml:

pip install .

or you can do it manually:

pip install PyQt6 opencv-python mediapipe numpy pygame

Usage

Open terminal or cd into pyqt/ folder then run the pyqt app:

python main.py

This will open the application

Dependencies

OpenCV – Real-time camera use

MediaPipe – Human pose detection

PyQT – UI framework

NumPy

PyGame – Audio handling

matplotlib

• Statistics visualization

todo:

• improve readme
• executable for simple use of app
• MORE statistical analysis oer longer periods
• improved error handling
• improved UI, allow for adjustable window size
• make threshold on stats page update with calibration