Skip to content

Track user's head movements with tensorflow's BlazeFace model to control a series of layered images with p5js and WebGL

Notifications You must be signed in to change notification settings

nsitu/2D-mirror

Repository files navigation

2D-mirror

Track user's head movements with tensorflow's BlazeFace model to control a series of layered images with p5js and WebGL

Demo

View a live demo at nsitu.ca/2dmirror

Background

This sketch tracks the user's face via webcam. Head movements are used as a kind of UI / control. A BlazeFace model estimates facial landmarks in webcam video. This sketch uses tensorflow's "wasm" (web assembly) backend (WebGL and CPU are also possible). Facial orientation / proximity is calculated from eye and nose locations. Layered images react dynamically to users' head movements.

Themes

Images in this iteration are based on a taxonomy by Giorgia Lupi. See also: Accurat's work with Data Portraits

Resources

Tensorflow BlazeFace - Landmarks Demo
BlazeFace Model Card - PDF

About

Track user's head movements with tensorflow's BlazeFace model to control a series of layered images with p5js and WebGL

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published