Mixed-reality snapshot system using environmental depth sensors

This paper proposes a method to merge the real and virtual world in snapshots captured by a mobile camera. While snapshots are usually captured in a dynamic scene that temporally changes the appearance, we developed a pose estimation method for a mobile camera by using depth sensors installed in the capturing environment. In order to generate a fine MR (Mixed-Reality) image, our method segments overlapping or overlapped regions by combining a graph-cut technique and captured depth data. A pilot system is presented to confirm the effectiveness of our proposed methods.