We consider a face-to-face\nvideoconferencing system that uses a Kinect camera at\neach end of the link for 3D modeling and an ordinary\n2D display for output. The Kinect camera allows\na 3D model of each participant to be transmitted;\nthe (assumed static) background is sent separately.\nFurthermore, the Kinect tracks the receiver�s head,\nallowing our system to render a view of the sender\ndepending on the receiver�s viewpoint. The resulting\nmotion parallax gives the receivers a strong impression\nof 3D viewing as they move, yet the system only needs\nan ordinary 2D display. This is cheaper than a full\n3D system, and avoids disadvantages such as the need\nto wear shutter glasses, VR headsets, or to sit in a\nparticular position required by an autostereo display.\nPerceptual studies show that users experience a greater\nsensation of depth with our system compared to a\ntypical 2D videoconferencing system.
Loading....