Authors:
Rajat Sharma
1
;
Tobias Schwandt
1
;
Christian Kunert
1
;
Steffen Urban
2
and
Wolfgang Broll
1
Affiliations:
1
Ilmenau University of Technology, Virtual Worlds and Digital Games Group, Ehrenbergstraße 29, Ilmenau, Germany
;
2
Carl Zeiss AG, Corporate Research and Technology, Carl-Zeiss-Promenade 10, Jena, Germany
Keyword(s):
Point Cloud Upsampling, Surface Normal Estimation, Surface Reconstruction, Deep Learning.
Abstract:
The reconstruction of real-world surfaces is on high demand in various applications. Most existing reconstruction approaches apply 3D scanners for creating point clouds which are generally sparse and of low density. These points clouds will be triangulated and used for visualization in combination with surface normals estimated by geometrical approaches. However, the quality of the reconstruction depends on the density of the point cloud and the estimation of the surface normals. In this paper, we present a novel deep learning architecture for point cloud upsampling that enables subsequent stable and smooth surface reconstruction. A noisy point cloud of low density with corresponding point normals is used to estimate a point cloud with higher density and appendant point normals. To this end, we propose a compound loss function that encourages the network to estimate points that lie on a surface including normals accurately predicting the orientation of the surface. Our results show t
he benefit of estimating normals together with point positions. The resulting point cloud is smoother, more complete, and the final surface reconstruction is much closer to ground truth.
(More)