Pose Estimation with the Fastest Python Deep Learning Model (MoveNet Lightning) (detect 17 joints) just you should set the address of input images which contains human. I decided to change this code to predict 32 joint of input human image
I learn this code from this link(but he's set webcam as an input): https://youtu.be/SSW9LzOJSus