webgazer
Version:
WebGazer.js is an eye tracking library that uses common webcams to infer the eye-gaze locations of web visitors on a page in real time. The eye tracking model it contains self-calibrates by watching web visitors interact with the web page and trains a map
30 lines (27 loc) • 910 B
JavaScript
const params = {
moveTickSize: 50,
videoContainerId: 'webgazerVideoContainer',
videoElementId: 'webgazerVideoFeed',
videoElementCanvasId: 'webgazerVideoCanvas',
faceOverlayId: 'webgazerFaceOverlay',
faceFeedbackBoxId: 'webgazerFaceFeedbackBox',
gazeDotId: 'webgazerGazeDot',
videoViewerWidth: 320,
videoViewerHeight: 240,
faceFeedbackBoxRatio: 0.66,
// View options
showVideo: true,
mirrorVideo: true,
showFaceOverlay: true,
showFaceFeedbackBox: true,
showGazeDot: true,
camConstraints: { video: { width: { min: 320, ideal: 640, max: 1920 }, height: { min: 240, ideal: 480, max: 1080 }, facingMode: "user" } },
dataTimestep: 50,
showVideoPreview: true,
applyKalmanFilter: true,
saveDataAcrossSessions: true,
// Whether or not to store accuracy eigenValues, used by the calibration example file
storingPoints: false,
trackEye: 'both',
};
export default params;