情報処理学会 第81回全国大会講演要旨

The traditional object recognition AR frameworks are mostly based on the image recognition and are not quite useful for recognizing 3D printed objects and their pose estimation because they only has single color generally. In this paper, the authors are trying to build a new object recognition AR framework that deals with the depth-map of 3D printed object captured using Kinect sensor to recognize its pose, position and orientation, using Point Cloud Library. This framework will enable to build the new type of AR based educaitonal materials in the near future.