Machine Learning: ML5 and PoseNet - teaching computers human gestures

I have been waiting to learn about and experience more of the big topic of the last few years - Machine Learning. It was such a fun learning experience, teaching the computer to recognize myself, objects, facial expressions, visually, using a machine learning library called ML5 developed by ITP. One of the featured libraries within ML5 is PoseNet - powered by TensorFlow.js - a project by Google Creative Lab, with contribution from our Dan Oved, an ITP second year student. Below is the link to experience my first machine learning project for this week. I chose to play around with PoseNet which captures Real-time Human Pose Estimation via webcam.

https://editor.p5js.org/son/full/BypW6AJTX