SlidAR: A 3D positioning method for SLAM-based handheld augmented reality

J. Polvi , T. Taketomi , G. Yamamoto , A. Dey , C. Sandor , H. Kato

Comput. Graph.

Placing virtual objects accurately in handheld AR is hard without markers or a full 3D map of the scene. SlidAR positions objects for SLAM-based handheld AR using 3D ray-casting and epipolar geometry, needing neither a perfect reconstruction of the environment nor virtual depth cues. In a user study against a device-centric method the authors call HoldAR, SlidAR was significantly faster, required less device movement, and scored better subjectively, with somewhat higher accuracy.

@article{polvi2016,
  author    = {J. Polvi and T. Taketomi and G. Yamamoto and A. Dey and C. Sandor and H. Kato},
  title     = {SlidAR: A 3D positioning method for SLAM-based handheld augmented reality},
  journal = {Comput. Graph.},
  volume    = {55},
  pages     = {33-43},
  doi       = {10.1016/j.cag.2015.10.013},
  year      = {2016}
}