Skip to content
/ AIgym Public

Fitness application with personal AI coach (on Mediapipe and OpenCV).

License

Notifications You must be signed in to change notification settings

raulgad/AIgym

Repository files navigation

AIgym - fitness application with personal AI coach

AIgym was created in order to practise yoga and workout in real time using computer vision

  • AIgym supports 2 modes: yoga and workout

  • There are 3 exercises in each program:

    Yoga: T-Pose, Warrior Pose, Tree Pose

    Workout: Arms, Legs, Push-ups

  • AIgym was created using MediaPipe, OpenCV, PyInstaller.

The example of pose handling

process

Remote hand control

virtual_control

The process of training consists of the following steps:

  • Recieve video stream using OpenCV.
  • Video frames are passed to MediaPipe Pose model that detects pose, adds landmarks and records their coordinates.
  • Landmarks coordinates are extracted, organized and passed to the model for prediction.
  • The total time of the entire training is displayed in the upper left corner within the "Pause" button (see the gifs above).
  • Yoga pose timing and workout repetitions (that user must accomplish) are displayed in the upper right corner within the "Next" button (see the gifs above).
  • Incorrect pose indication through user's body lines.

This project was completed in 10 days.

About

Fitness application with personal AI coach (on Mediapipe and OpenCV).

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages