Real-Time Occlusion Culling for Augmented Reality

This paper proposes a real-time occlusion culling for augmented reality(AR). In AR applications, it is difficult to realistically overlay a virtual object when it should be partially occluded by real objects. One way of handling this occlusion is to use the Z-keying method which requires a dense depth map. However, the accuracy of depth map in real-time is far lower than the required one even with the state-of-the-art technology. Therefore, this paper proposes a method of occlusion culling exploiting foreground silhouette and depth value obtained by an efficient region of interest(ROI) extraction method from background subtraction(BS) and stereo depth(SD) algorithms. Experimental results show that the proposed method can produce composited images of reasonable quality and thus it can be successfully applied to AR applications.This work was partly supported by the IT R&D program of MKE/IITA. [2009-F-208-01, Signal Processing Elements and their SoC Developments to Realize the Integrated Service System for Interactive Digital Holograms