Workout form analysis model preview

Stress Analysis App

I developed a computer vision–based stress monitoring system that analyzes physiological signals directly from a live camera feed. Using pose estimation and facial landmark detection, the application tracks blink rate and shoulder movement in real time to help users analyze their breathing behavior.

What it does

  • Guides inhale phase (4 seconds)
  • Guides breath hold (4 seconds)
  • Guides exhale phase (4 seconds)
  • Guides final hold (4 seconds)
  • Measures blink count and completed breathing cycles

If the user deviates from the 4–4–4–4 rhythm, the system suggests repeating the round. If the breathing pattern matches the expected cycle, it confirms successful completion.

Project overview

A detailed explanation of the whole process of creating this system. It includes codes, screenshots, and descriptions.

Read the article

Used technology

  • Python
  • MediaPipe
  • OpenCV
  • Pandas

Status

The demo version is done.

Project Demo