Overview
Explore pose estimation using ml5.js in this 21-minute video tutorial. Learn to implement the BodyPose model with MoveNet and BlazePose, tracking body keypoints and visualizing 'skeleton' connections using live video. Dive into coding examples, starting with single image detection and progressing to more complex applications like adding a virtual hat. Discover the underlying concepts, including the COCO dataset and creative inspirations. Follow along with provided p5.js Web Editor sketches to practice detecting specific keypoints, creating skeleton connections, and even experimenting with 3D BlazePose. Gain insights into machine learning applications for webcam-based pose detection, equipping you with skills to create interactive and engaging projects using JavaScript and ml5.js.
Syllabus
Introduction
What is the BodyPose model?
Coco Dataset
Creative Inspiration
Start coding
Working with specific keypoints
Skeleton connections between the keypoints
Using dist function
BlazePose
Outro
Taught by
The Coding Train