the-midnight-paper

RetroDepth: 3D Silhouette Sensing for High-Precision Input On and Above Physical Surfaces

Authors: David Kim, Shahram Izadi, Jakub Dostal, Christoph Rhemann, Cem Keskin, Christopher Zach, Jamie Shotton, Tim Large, Steven Bathiche, Matthias Niebner, D. Alex Butler, Sean Fanello, Vivek Pradeep

CHI 2014, April 26 - May 01 2014, Toronto, ON, Canada

Keywords: NUI, 3D contours, depth sensing, 3D input, touch, stylus, vision-based UIs, stereo matching, contour classification

Strength

RetroDepth senses only the 3D silhouettes of interacting objects which greatly lowers the computation required for depth sensing. Due to this fast and precise 3D input precision, RetroDepth support all common input modalities of touch, pressure and hover in real-time.

Weakness

The accuracy of RetroDepth is dependent on the use of retro-reflective materials to get a uniform response. While this gives great results, it cannot be extended for in-air gesture detection where there is no retro-reflective surface in the background.

Future Work

It would be interesting to see how RetroDepth’s 3D input precision varies if there are no retro-reflective surface in the background.