Natural Feature Tracking in JavaScript

Abstract

Computer Vision-based natural feature tracking is at the core of modern Augmented Reality applications. Still, Web-based Augmented Reality typically relies on location-based sensing (using GPS and orientation sensors) or marker-based approaches to solve the pose estimation problem.
In our research, we have investigated two possibilities to bring efficient Natural Feature Tracking to Standard Web-Browsers.

First, we implemented and evaluated an efficient natural feature tracking pipeline for standard Web browsers using HTML5 and WebAssembly. Our system can track image targets at real-time frame rates on tablet PCs (up to 60 Hz) and smartphones (up to 25 Hz).

You can find a live demonstration of this system here. For the live demonstration to work, you need to print out this target first.


In previous work, we developed a web-based Augmented Reality system built only on JavaScript, WebGL and HTML5. At the core of our system is an efficient Natural Feature Tracking pipeline solely implemented in JavaScript. We demonstrate real-time framerates on desktop computers and interactive framerates on smartphones.

Media



Publications

  • Fabian Göttl, Philipp Gagel & Jens Grubert. Efficient Pose Tracking from Natural Features in Standard Web Browsers. In Proceedings of the 23rd International ACM Conference on 3D Web Technology (Web3D). 20-22 June 2018, Poznań, Poland (to appear). arxiv preprint | doi
  • Christoph Oberhofer, Jens Grubert & Gerhard Reitmayr. Natural Feature Tracking in JavaScript. In Proceedings of the IEEE Virtual Reality Conference 2012 (IEEE VR 2012). March 04-08, 2012, Orange County, California, USA (pp. 111-112). pdf | poster