Skip to main content
100%

three-vrm example

✓ Published0🌍 Public
HHirosaji
Last edited Feb 25, 2020
Created on Nov 25, 2019

A 3D VRM avatar performs motion capture animation synced to a live webcam feed. The webcam video and PoseNet-detected keypoints are drawn onto a 2D overlay canvas, while the estimated pose data drives the avatar’s joint rotations in a Three.js WebGL scene with GridHelper and AxesHelper. The code uses TensorFlow.js PoseNet, THREE.VRM, GLTFLoader, and OrbitControls, rendering the avatar from a local .vrm model.

AI-generated description

▶︎ What's this?: RM motion capture demo using pixiv three-vrm

▶︎ Reference: https://github.com/pixiv/three-vrm


[Gist] https://gist.github.com/Hirosaji/00d3c9ceba65ef7ddf0f2890968132dc


Built with blockbuilder.org

mit Licensed

Similar vizzes