> we are far from photo realistic human representations on screen, let alone VR.
Would we actually need to simulate human representations in order for it to be photorealistic? For example, what if your VR headset displays information consumed from a set of stereoscopic cameras?
There would be a lot of complications. (Won't the other guy also be wearing a VR headset, thus obscuring the cameras? Though not necessarily if it's one-directional)
Imagine a VR conversation with another person who's in front of a stereo camera. You can't move your headset much, but you can tilt and pan slightly - basically stay in place and move your head. Imagine combining the VR technology with 3d scene reconstruction such as: http://petapixel.com/2013/07/24/researchers-reconstruct-high... - performed in realtime using the stereo camera. It could potentially present a fairly convincing VR experience of speaking with them in person.
(I assume that you need to perform scene reconstruction in order to generate a display according to the user's head movements in his VR headset given that the camera can't track and mimic the movement in realtime.)
> we are far from photo realistic human representations on screen, let alone VR.
Would we actually need to simulate human representations in order for it to be photorealistic? For example, what if your VR headset displays information consumed from a set of stereoscopic cameras?
There would be a lot of complications. (Won't the other guy also be wearing a VR headset, thus obscuring the cameras? Though not necessarily if it's one-directional)
Imagine a VR conversation with another person who's in front of a stereo camera. You can't move your headset much, but you can tilt and pan slightly - basically stay in place and move your head. Imagine combining the VR technology with 3d scene reconstruction such as: http://petapixel.com/2013/07/24/researchers-reconstruct-high... - performed in realtime using the stereo camera. It could potentially present a fairly convincing VR experience of speaking with them in person.
(I assume that you need to perform scene reconstruction in order to generate a display according to the user's head movements in his VR headset given that the camera can't track and mimic the movement in realtime.)