6-DoF Pose Estimation and CAD Model Retrieval for XR Interface from a Single RGB Image

This paper proposes a 6-DoF pose estimation and CAD model retrieval for XR interface from a single RGB image. A deep learning network is used to estimate the 6-DoF pose of a real object in an RGB image. Then, several CAD model candidates are rendered and their similarity with the image is measured to decide the final matching CAD model. In the last, the rendered image of the matching CAD model is overlayed with the RGB image. In the experiment, total ten object categories are tested and evaluated using two deep networks.