Common 3D acquisition techniques, such as laser scanning and stereo capture, are realistically only 2.5D in nature. Here, we consider the automated completion of hidden or missing portions in 3D scenes originally acquired from 2.5D (or 3D) capture. We propose an approach based on the nonparametric propagation of available scene knowledge from the known (visible) scene areas to these unknown (invisible) 3D regions in conjunction with an initial underlying geometric surface completion.
Toby P. Breckon, Robert B. Fisher