CARMa: Content augmented reality marker

Abstract

The current marker-based augmented reality (AR) rendering has demonstrated good results for online and special purpose applications such as computer-assisted tasks and virtual training. However, it fails to deliver a solution for off-line and generic applications such as augmented books, newspapers, and scientific articles. These applications feature too many markers that imposes a serious challenge on the recognition module. This paper proposes a novel design for augmented reality markers. The proposed marker design employs multi-view orthographic projection to derive dense depth maps and relies on splats rendering for visualisation. The main objective is to interpret the marker rather than recognising it. The proposed marker design stores six depth map projections of the 3D model along with their coloured textures in the marker.