Turning My Body Into a Controller with Machine Learning, ml5.js, and p5.js

Subscribers:
6,610
Published on ● Video Link: https://www.youtube.com/watch?v=96sWFP9CCkQ



Duration: 9:42
1,396 views
39


In this video we look at some applications from the ml5.js site (and also my own drawing program) and it's use in p5.js (Javascript) using machine learning pre-trained models to track body movements - hand, face, and full body tracking. You might use this to move an object around the screen using any body part. Or train the computer to recognize that a certain pose means something. We look at PoseNet, Handpose, Facemesh, and KNNClassifier PoseNet, as well as Dan Shiffman's PoseNet YMCA example. ml5.js is built around TensorFlow.js.

My drawing program code:
https://editor.p5js.org/StevesMakerspace/sketches/VCFXH2lOh

All about ml5.js:
https://ml5js.org/

Dan Shiffman's Coding Train video usine PoseNet:
https://www.youtube.com/watch?v=FYgYyq-xqAw

Join my Discord: https://discord.gg/FSW6zzU

Follow me on Twitter: https://twitter.com/SteveMakerspace

Music:
Funk Down (Sting) by MK2 (theme)
Come and Get It! by Dan Lebowitz
Blues Infusion by Quincas Moreira
Dewey, Cheedham, and Howe by John Deley and the 41 Players
Hold on a Second by John Deley and the 41 Players







Tags:
Tensorflow.js
tensorflow
machine learning
neural network
p5.js
ml5.js
posenet
facemesh
handpose
KNNClassifier
coding train
dan shiffman
javascript machine learning
p5.js machine learning
processing machine learning
pre-trained model