As the mobile phone hardware platform evolves, 2D barcode scanning with camera phones have emerged as one of killer mobile applications. In our context, we consider it as a registration target, where virtual contents are synthesized in real-time.

We propose a real-time 2D barcode detection method for augmented reality applications. Our method embeds a set of information about the specifications of a barcode and exploits the embedded information for detection and tracking a barcode. Our method combines natural feature tracking to allow barcode tracking under partial occlusions. 6 DoF pose of the barcode is estimated and virtual contents are stably synthesized in real-time video on mobile phones.