1

Our graduation project is to track human motion at real time while doing an exercise and then give him feedback if that exercise is right or wrong and instructions to make it a perfect one, so we split the code into two parts. The first one is estimating all human keypoints (joint) and that happen by the repo here by using tensorflow : https://github.com/ildoonet/tf-pose-estimation and we run this code, second part is to perform on a small exercise like bind the arm (hand part) 90 degree toward the shoulder , so we need to detect the angles between limbs and give instructions (voice or written commands) according to this angle, so anyone could help with an idea or a known code ?

Lisandro Acosta
  • 3
  • 2
Aya N. Mohammed
  • 31
  • 6

1 Answers1

1

In module tf_pose.estimator there is class Human, list of which is returned from method PoseEstimator.inference(). From Human.body_parts you can get all the positions and connections - joints. Then you can compute the angles between parts in joints. Should be some quite easy trigonometry with math.atan2 function to get angle absolute to positive x-axis. Then you can compute how are the agles relative to each other in a link, starting from shoulder for example.

karna7
  • 176
  • 7