Gaze tracking directly in the browser
Use any standard webcam
9 point calibration system
Coded in vanilla JavaScript
Open-Source and Developer friendly documentation
Built by Computer Science students at The University of Auckland in 12 weeks
Peeper.js uses OpenCV 4.5.0 and Face-api 0.22.2 with the Tiny Face Detector and 68 Point Face Landmark Detection.
Download the example.zip release file here.
Navigate to the example folder. This folder contains the example HTML page, required modules/libraries, a peeper.js build file and a main.js file which calls relevant functions in the correct order. This folder is independent of the other files in the repo and can be moved to your desired location.
From here you can simply run a live server to view the example page in your browser.
This project was coded in under 12 weeks by undergraduate students of The University of Auckland. Although we managed to create a viable prototype in the eyes of our client, there are a few limitations of peeper.js. Some of these include:
NOTE: This project is a fork of the original
Peeper.js was developed by the students from the University of Auckland in the context of a capstone project with the aim of creating an open-source gaze-tracking library. The project task was proposed by Dr. Gerald Weber and was implemented based on different open source libraries and studies found on the web such as:
