Pose Trainer: Correcting Exercise Pose using Pose Estimation

Steven Chen and Richard Yang

Read the paper here: https://www.researchgate.net/publication/324759769_Pose_Trainer_Correcting_Exercise_Posture_using_Pose_Estimation

---

WARNING: This code is experimental, and is no longer maintained or supported. It is not expected to work.

The paper is a much better resource than this repository for understanding how Pose Trainer works.

It's highly recommended to use this repository only as a reference for learning, not as a working program to build upon.

---

Setup notes (Please read the warning above):

Please follow the instructions at https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/installation/0_index.md#windows-portable-demo to download and set up OpenPose, and move it into the root folder of the repository. (ex. If this repository is x/y/z/pose-trainer, then OpenPose should be placed at x/y/z/pose-trainer/openpose.)

Pose trainer has been tested with OpenPose 1.7.0.

• Pose Trainer is designed to run on Windows only, since it uses the Windows portable version of OpenPose.
• Pose Trainer expects the OpenPose folder to be in Pose Trainer's repository folder.
• Pose Trainer should be run from the root folder of the repository using main.py.
• Pose Trainer requires Python with the necessary libraries installed.

main.py is the primary script for running Pose Trainer. For options: py main.py --help

Sample command: py main.py --mode evaluate --video sample_bicep_curl.mp4

About speed:

• OpenPose can run on a CPU-only machine, but it will be very slow.
• If you have a computer with an NVIDIA GPU, OpenPose will run significantly faster.